{ "cells": [ { "cell_type": "markdown", "id": "28f36efa-b9f8-4935-a403-a4010cdbad8e", "metadata": {}, "source": [ " \"Header\" " ] }, { "cell_type": "markdown", "id": "fa0ee592-705e-49c5-a26c-dd7cacfbb545", "metadata": {}, "source": [ "# Computer Vision for Industrial Inspection #" ] }, { "cell_type": "markdown", "id": "17a48909-8799-4b25-bd11-91cc92fb46a1", "metadata": { "jp-MarkdownHeadingCollapsed": true }, "source": [ "## Part 2 - Model Training with Transfer Learning ##\n", "In this notebook, you will learn how to train a classification model with the TAO Toolkit using pre-trained ResNet50 weights. \n", "\n", "**Table of Contents**\n", "
\n", "This notebook covers the below sections: \n", "1. [Introduction to TAO Toolkit](#s2-1)\n", " * [Transfer Learning](#s2-1.1)\n", " * [Vision AI Pre-Trained Models Supported](#s2-1.2)\n", " * [TAO Toolkit Workflow](#s2-1.3)\n", " * [TAO Launcher, CLI (Command Line Interface), and Spec Files](#s2-1.4)\n", " * [Set Up Environment Variables](#s2-1.5)\n", " * [Exercise #1 - Explore TAO Toolkit CLI](#s2-e1)\n", "2. [Training a Classification Model](#s2-2)\n", " * [Preparation for Model Training](#s2-2.1)\n", " * [Download Pre-Trained Model](#s2-2.2)\n", " * [Prepare Dataset](#s2-2.3)\n", " * [Model Training](#s2-2.4)\n", " * [Exercise #2 - Modify Model Config](#s2-e2)\n", " * [Exercise #3 - Modify Train Config](#s2-e3)\n", " * [Combine Configuration Files and Initiate Model Training](#s2-2.5)\n", " * [Model Evaluation](#s2-2.6)\n", " * [Exercise #4 - Modify Eval Config](#s2-e4)\n", " * [Combine Configuration Files and Evaluate Model](#s2-2.7)\n", "4. [Model Tuning](#s2-3)\n", " * [Data Augmentation](#s2-3.1)\n", " * [Exercise #5 - Retrain Model](#s2-e5)\n", " * [Model Inference](#s2-3.2)\n", " * [Precision and Recall](#s2-3.3)\n", " * [Adjusting the Threshold](#s2-3.4)" ] }, { "cell_type": "markdown", "id": "c4ae3a29-df70-4396-a6f4-d1d8d164d5ff", "metadata": {}, "source": [ "\n", "## Introduction to the TAO Toolkit ##\n", "The TAO Toolkit, Train Adapt Optimize, is a framework that simplifies the AI/ML model development workflow. It lets developers fine-tune pre-trained models with custom data to produce highly accurate computer vision models efficiently, eliminating the need for large training runs and deep AI expertise. In addition, it also enables model optimization for inference performance. You can learn more about the TAO Toolkit [here](https://developer.nvidia.com/tao-toolkit) or read the documentation [here](https://docs.nvidia.com/tao/tao-toolkit/index.html#). \n", "

\n", "\n", "The TAO Toolkit uses pre-trained models to accelerate the AI development process and reduce costs associated with large scale data collection, labeling, and training models from scratch. Transfer learning with pre-trained models can be used for classification, object detection, and image segmentation tasks. The TAO Toolkit offers useful features such as: \n", "* Low-coding approach that requires no AI framework expertise, reducing the barrier of entry for anyone who wants to get started building AI-based applications\n", "* Flexible configurations that allow customization to help advance users prototype faster\n", "* Large catalogue of production-ready pre-trained models for common CV tasks that can also be customized with users' own data\n", "* Easy to use interface for model optimization such as pruning and quantization-aware training\n", "* Integration with the Triton Inference Server\n", "\n", "_Note: The TAO Toolkit comes with a set of reference scripts and configuration specifications with default parameter values that enable developers to kick-start training and fine-tuning. This lowers the bar and enables users without a deep understanding of models, expertise in deep learning, or beginning coding skills to be able to train new models and fine-tune the pretrained ones._" ] }, { "cell_type": "markdown", "id": "ad8aca25-bc7a-4dc7-a4d3-b903cbdf13f4", "metadata": {}, "source": [ "\n", "### Transfer Learning ###\n", "In practice, it is rare and inefficient to initiate the learning task on a network with randomly initialized weights due to factors like data scarcity (inadequate number of training samples) or prolonged training times. One of the most common techniques to overcome this is to use transfer learning. Transfer learning is the process of transferring learned features from one application to another. It is a commonly used training technique where developers use a model trained on one task and re-train to use it on a different task. This works surprisingly well as many of the early layers in a neural network are the same for similar tasks. For example, many of the early layers in a convolutional neural network used for a Computer Vision (CV) model are primarily used to identify outlines, curves, and other features in an image. The network formed by these layers are referred to as the **backbone** of a more complex model. Also known as feature extractors, they take as input the image and extracts the feature map upon which the rest of the network is based. The learned features from these layers can be applied to similar tasks carrying out the same identification in other domains. Transfer learning enables adaptation (fine-tuning) of an existing neural network to a new one, which requires significantly less domain-specific data. In most cases, fine-tuning takes significantly less time (a reduction by x10 factor is common), saving time and resources. As it relates to vision AI, transfer learning can be used for adding new classification by transferring weights from one application to another. \n", "\n", "

\n", "\n", "More information about transfer learning can be found in this [blog](https://blogs.nvidia.com/blog/2019/02/07/what-is-transfer-learning/)." ] }, { "cell_type": "markdown", "id": "afb01061-8a05-4c4f-96aa-28926278b503", "metadata": {}, "source": [ "\n", "### Vision AI Pre-Trained Models Supported ###\n", "Developers, system builders, and software partners building optical inspection systems can bring their own custom data to train with and fine-tune pre-trained models quickly instead of going through significant effort in large data collection and training from scratch. **General purpose vision models** provide pre-trained weights for popular network architectures to train an image classification model, an object detection model, or a segmentation model. This gives users the flexibility and control to build AI models for any number of applications, from smaller lightweight models for edge deployment to larger models for more complex tasks. They are trained on the [Open Images](https://opensource.google/projects/open-images-dataset) dataset and provide a much better starting point for training versus training from scratch or starting from random weights. \n", "\n", "The TAO Toolkit adapts popular network architectures and backbones to custom data, allowing developers to train, fine tune, prune, and export highly optimized and accurate AI models. When working with TAO, first choose the model architecture to be built, then choose one of the supported backbones. \n", "

\n", "\n", "_Note: The pre-trained weights from each feature extraction network merely act as a starting point and may not be used without re-training. In addition, the pre-trained weights are network specific and shouldn't be shared across models that use different architectures._" ] }, { "cell_type": "markdown", "id": "0319ca49-760b-4aa1-b2f5-8c31657c125c", "metadata": {}, "source": [ "\n", "### TAO Toolkit Workflow ###\n", "Building optical inspection systems is hard. And tailoring even a single component to the needs of the enterprise for deployment is even harder. Deployment for a domain-specific application typically requires several cycles of re-training, fine-tuning, and deploying the model until it satisfies the requirements. As a starting point, model development with the TAO Toolkit typically follows the below steps: \n", "\n", "0. Configuration\n", "1. Download a pre-trained model from NGC\n", "2. Prepare the data for training\n", "3. Train the model using transfer learning\n", "4. Evaluate the model for target predictions\n", "5. Export the model\n", "* Steps to optimize the model for improved inference performance\n", "\n", "

" ] }, { "cell_type": "markdown", "id": "bcb69288-90df-455f-b53a-0a32adf74dbc", "metadata": {}, "source": [ "\n", "### TAO Launcher, CLI (Command Line Interface), and Spec Files ###\n", "The TAO Toolkit is a low-coding framework that makes it easy to get started. It uses a **launcher** to pull from NGC registry and instantiate the appropriate TAO container that performs the desired subtasks such as convert data, train, evaluate, or export. The TAO launcher is a python package distributed as a python wheel listed in the `nvidia-pyindex` python index, which has been prepared for you already. \n", "\n", "Users interact with the launcher with its **Command Line Interface** that is configured using simple [**Protocol Buffer**](https://developers.google.com/protocol-buffers) **specification files** to include parameters such as the dataset parameters, model parameters, and optimizer and training hyperparameters. More information about the TAO Toolkit Launcher can be found in the [TAO Docs](https://docs.nvidia.com/tao/tao-toolkit/text/tao_launcher.html#tao-launcher). \n", "\n", "The tasks can be invoked from the TAO Toolkit Launcher using the convention `tao `, where `` are the arguments required for a given subtask. Once the container is launched, the subtasks are run by the TAO Toolkit containers using the appropriate hardware resources. \n", "

" ] }, { "cell_type": "markdown", "id": "cdedcccb-f6d7-4394-84b1-94e008f8e884", "metadata": {}, "source": [ "

\n", "Since the TAO Toolkit uses the launcher to pull containers, the first time running a task may take extra time to load. " ] }, { "cell_type": "markdown", "id": "65f7f4ec-1826-499f-b001-28d1eb0d71a1", "metadata": {}, "source": [ "\n", "### Set Up Environment Variables ###\n", "We set up a couple of environment variables to help us mount the local directories to the TAO container. Specifically, we want to set paths for the `$LOCAL_TRAINING_DATA`, `$LOCAL_SPEC_DIR`, and `$LOCAL_PROJECT_DIR` for the output of the TAO experiment with their respective paths in the TAO container. In doing so, we can make sure that the TAO experiment generated collaterals such as checkpoints, model files (e.g. `.hdf5`, `.tlt`, or `.etlt`), and logs are output to `$LOCAL_PROJECT_DIR/classification`. \n", "\n", "_Note that users will be able to define their own export encryption key when training from a general-purpose model. This is to protect proprietary IP and used to decrypt the `.etlt` model during deployment._" ] }, { "cell_type": "code", "execution_count": 1, "id": "446943db-e633-4b5b-9997-c767a849085d", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "env: KEY=my_model_key\n", "env: LOCAL_PROJECT_DIR=/dli/task/tao_project\n", "env: LOCAL_DATA_DIR=/dli/task/tao_project/data\n", "env: LOCAL_SPECS_DIR=/dli/task/tao_project/spec_files\n", "env: TAO_PROJECT_DIR=/workspace/tao-experiments\n", "env: TAO_DATA_DIR=/workspace/tao-experiments/data\n", "env: TAO_SPECS_DIR=/workspace/tao-experiments/spec_files\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# set environment variables\n", "import os\n", "import pandas as pd\n", "import time\n", "import shutil\n", "import json\n", "import numpy as np\n", "import matplotlib.image as mpimg\n", "import matplotlib.pyplot as plt\n", "import math\n", "\n", "%set_env KEY=my_model_key\n", "\n", "%set_env LOCAL_PROJECT_DIR=/dli/task/tao_project\n", "%set_env LOCAL_DATA_DIR=/dli/task/tao_project/data\n", "%set_env LOCAL_SPECS_DIR=/dli/task/tao_project/spec_files\n", "os.environ[\"LOCAL_EXPERIMENT_DIR\"]=os.path.join(os.getenv(\"LOCAL_PROJECT_DIR\"), \"classification\")\n", "\n", "%set_env TAO_PROJECT_DIR=/workspace/tao-experiments\n", "%set_env TAO_DATA_DIR=/workspace/tao-experiments/data\n", "%set_env TAO_SPECS_DIR=/workspace/tao-experiments/spec_files\n", "os.environ['TAO_EXPERIMENT_DIR']=os.path.join(os.getenv(\"TAO_PROJECT_DIR\"), \"classification\")\n", "\n", "# make the data directory\n", "!mkdir -p $LOCAL_DATA_DIR\n", "\n", "# unzip\n", "!unzip -qq data/viz_BYD_new.zip -d data\n", "\n", "# remove zip file\n", "!rm data/viz_BYD_new.zip" ] }, { "cell_type": "markdown", "id": "abe38326-8dc7-438f-b392-913166d3a185", "metadata": {}, "source": [ "The cell below maps the project directory on your local host to a workspace directory in the TAO docker instance, so that the data and the results are mapped from in and out of the docker. This is done by creating a `.tao_mounts.json` file. For more information, please refer to the [launcher instance](https://docs.nvidia.com/tao/tao-toolkit/tao_launcher.html) in the user guide. Setting the `DockerOptions` ensures that you don't have permission issues when writing data into folders created by the TAO docker." ] }, { "cell_type": "code", "execution_count": 2, "id": "e0a30e1d-a66d-4752-85d7-6676f31c3431", "metadata": {}, "outputs": [], "source": [ "# DO NOT CHANGE THIS CELL\n", "# mapping up the local directories to the TAO docker\n", "mounts_file = os.path.expanduser(\"~/.tao_mounts.json\")\n", "\n", "drive_map = {\n", " \"Mounts\": [\n", " # Mapping the project directory\n", " {\n", " \"source\": os.environ[\"LOCAL_PROJECT_DIR\"],\n", " \"destination\": \"/workspace/tao-experiments\"\n", " },\n", " # Mapping the specs directory.\n", " {\n", " \"source\": os.environ[\"LOCAL_SPECS_DIR\"],\n", " \"destination\": os.environ[\"TAO_SPECS_DIR\"]\n", " },\n", " # Mapping the data directory.\n", " {\n", " \"source\": os.environ[\"LOCAL_DATA_DIR\"],\n", " \"destination\": os.environ[\"TAO_DATA_DIR\"]\n", " },\n", " ],\n", " \"DockerOptions\": {\n", " \"user\": \"{}:{}\".format(os.getuid(), os.getgid())\n", " }\n", "}\n", "\n", "# writing the mounts file\n", "with open(mounts_file, \"w\") as mfile:\n", " json.dump(drive_map, mfile, indent=4)" ] }, { "cell_type": "markdown", "id": "0926eb3b-695f-41c7-b6ea-1372a0e3b7fb", "metadata": {}, "source": [ "To see the usage of different functionality that are supported, use the `-h` or `--help` option. For more information, see the [TAO Toolkit Quick Start Guide](https://docs.nvidia.com/tao/tao-toolkit/text/tao_toolkit_quick_start_guide.html). \n", "Here is the **sample output**: " ] }, { "cell_type": "code", "execution_count": 10, "id": "0d2c2ed9-6c06-4e74-aa9e-1dc3034228c9", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "usage: tao [-h] {list,stop,info,dataset,deploy,model} ...\n", "\n", "Launcher for TAO Toolkit.\n", "\n", "options:\n", " -h, --help show this help message and exit\n", "\n", "task_groups:\n", " {list,stop,info,dataset,deploy,model}\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "!tao -h" ] }, { "cell_type": "markdown", "id": "0a24d50a-f760-483c-90e0-680a45869b26", "metadata": {}, "source": [ "With the TAO Toolkit, users can train models for object detection, classification, segmentation, optical character recognition, facial landmark estimation, gaze estimation, and more. In TAO's terminology, these would be the **tasks**, which support **subtasks** such as `train`, `prune`, `evaluate`, `export`, etc. Each task/subtask requires different combinations of configuration files to accommodate for different parameters, such as the dataset parameters, model parameters, and optimizer and training hyperparameters. Part of what makes TAO Toolkit so easy to use is that most of those parameters are hidden away in the form of experiment specification files (spec files). They are detailed in the [documentation](https://docs.nvidia.com/tao/tao-toolkit/#tao-toolkit) for reference. It's very helpful to have these resources handy when working with the TAO Toolkit. In addition, there are several specific tasks that help with handling the launched commands. \n", "\n", "Below are the tasks available in the TAO Toolkit, organized by their respective computer vision objectives. We grayed out the tasks for Conversational AI as they are out of scope for this course. \n", "\n", "" ] }, { "cell_type": "markdown", "id": "9d81e20c-14fa-4b8f-a630-ed8d9a509b59", "metadata": {}, "source": [ "\n", "### Exercise #1 - Explore TAO Toolkit CLI ###\n", "Let's explore some TAO Toolkit tasks. \n", "\n", "**Instructions**:
\n", "* Modify the ``s only and execute the cell, choosing a task from options such as: `[classification_tf1, detectnet_v2, mask_rcnn, etc.]`, follow by a subtask from options such as: `[calibration_tensorfile, dataset_convert, evaluate, export, inference, prune, train]`. " ] }, { "cell_type": "code", "execution_count": 12, "id": "d1746850-d5d9-466c-ad58-fb54c561859a", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "2026-02-02 17:24:37,597 [TAO Toolkit] [INFO] root 160: Registry: ['nvcr.io']\n", "2026-02-02 17:24:37,686 [TAO Toolkit] [INFO] nvidia_tao_cli.components.instance_handler.local_instance 360: Running command in container: nvcr.io/nvidia/tao/tao-toolkit:5.0.0-tf1.15.5\n", "2026-02-02 17:24:37,697 [TAO Toolkit] [INFO] nvidia_tao_cli.components.docker_handler.docker_handler 301: Printing tty value True\n", "Using TensorFlow backend.\n", "2026-02-02 17:24:38.586200: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 17:24:38,637 [TAO Toolkit] [WARNING] tensorflow 40: Deprecation warnings have been disabled. Set TF_ENABLE_DEPRECATION_WARNINGS=1 to re-enable them.\n", "2026-02-02 17:24:39,665 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use sklearn by default. This improves performance in some cases. To enable sklearn export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 17:24:39,703 [TAO Toolkit] [WARNING] tensorflow 42: TensorFlow will not use Dask by default. This improves performance in some cases. To enable Dask export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 17:24:39,707 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use Pandas by default. This improves performance in some cases. To enable Pandas export the environment variable TF_ALLOW_IOLIBS=1.\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:150: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_hue(img, max_delta=10.0):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:173: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_saturation(img, max_shift):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:183: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_contrast(img, center, max_contrast_scale):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:192: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_shift(x_img, shift_stddev):\n", "2026-02-02 17:24:41.505021: I tensorflow/core/platform/profile_utils/cpu_utils.cc:109] CPU Frequency: 2499995000 Hz\n", "2026-02-02 17:24:41.505418: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x7dfc810 initialized for platform Host (this does not guarantee that XLA will be used). Devices:\n", "2026-02-02 17:24:41.505449: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version\n", "2026-02-02 17:24:41.506922: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcuda.so.1\n", "2026-02-02 17:24:41.698915: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:24:41.700950: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x7c17bc0 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:\n", "2026-02-02 17:24:41.700982: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Tesla T4, Compute Capability 7.5\n", "2026-02-02 17:24:41.701280: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:24:41.703162: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1674] Found device 0 with properties: \n", "name: Tesla T4 major: 7 minor: 5 memoryClockRate(GHz): 1.59\n", "pciBusID: 0000:00:1e.0\n", "2026-02-02 17:24:41.703224: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 17:24:41.703347: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcublas.so.12\n", "2026-02-02 17:24:41.705488: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcufft.so.11\n", "2026-02-02 17:24:41.705602: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcurand.so.10\n", "2026-02-02 17:24:41.709366: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcusolver.so.11\n", "2026-02-02 17:24:41.710531: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcusparse.so.12\n", "2026-02-02 17:24:41.710607: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudnn.so.8\n", "2026-02-02 17:24:41.710735: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:24:41.712693: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:24:41.714496: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1802] Adding visible gpu devices: 0\n", "2026-02-02 17:24:41.714545: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 17:24:41.723305: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1214] Device interconnect StreamExecutor with strength 1 edge matrix:\n", "2026-02-02 17:24:41.723332: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1220] 0 \n", "2026-02-02 17:24:41.723343: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1233] 0: N \n", "2026-02-02 17:24:41.723551: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:24:41.725533: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:24:41.727377: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1359] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 13496 MB memory) -> physical GPU (device: 0, name: Tesla T4, pci bus id: 0000:00:1e.0, compute capability: 7.5)\n", "usage: classification train [-h] [--num_processes NUM_PROCESSES] [--gpus GPUS]\n", " [--gpu_index GPU_INDEX [GPU_INDEX ...]]\n", " [--use_amp] [--log_file LOG_FILE] -e\n", " EXPERIMENT_SPEC_FILE -r RESULTS_DIR [-k KEY]\n", " [--init_epoch INIT_EPOCH] [-v] [-c CLASSMAP]\n", " {train,prune,inference,export,evaluate,calibration_tensorfile}\n", " ...\n", "\n", "optional arguments:\n", " -h, --help show this help message and exit\n", " --num_processes NUM_PROCESSES, -np NUM_PROCESSES\n", " The number of horovod child processes to be spawned.\n", " Default is -1(equal to --gpus).\n", " --gpus GPUS The number of GPUs to be used for the job.\n", " --gpu_index GPU_INDEX [GPU_INDEX ...]\n", " The indices of the GPU's to be used.\n", " --use_amp Flag to enable Auto Mixed Precision.\n", " --log_file LOG_FILE Path to the output log file.\n", " -e EXPERIMENT_SPEC_FILE, --experiment_spec_file EXPERIMENT_SPEC_FILE\n", " Path to the experiment spec file.\n", " -r RESULTS_DIR, --results_dir RESULTS_DIR\n", " Path to a folder where experiment outputs should be\n", " written.\n", " -k KEY, --key KEY Key to save or load a .tlt model.\n", " --init_epoch INIT_EPOCH\n", " Set resume epoch.\n", " -v, --verbose Include this flag in command line invocation for\n", " verbose logs.\n", " -c CLASSMAP, --classmap CLASSMAP\n", " Class map file to set the class indices of the model.\n", "\n", "tasks:\n", " {train,prune,inference,export,evaluate,calibration_tensorfile}\n", "2026-02-02 17:24:42,405 [TAO Toolkit] [INFO] nvidia_tao_cli.components.docker_handler.docker_handler 363: Stopping container.\n" ] } ], "source": [ "!tao model classification_tf1 train --help" ] }, { "cell_type": "markdown", "id": "55f19dad-30cb-4d12-a15d-7268743ab2b3", "metadata": {}, "source": [ "

\n", "\n", "Did you get the below error message? This is likely due to a bad NGC configuration. Please check the NGC CLI and Docker Registry section of the [introduction notebook](00_introduction.ipynb)." ] }, { "cell_type": "raw", "id": "ff968ed5-77ee-4237-8709-f55e96cc70b0", "metadata": {}, "source": [ "AssertionError: Config path must be a valid unix path. No file found at: /root/.docker/config.json. Did you run docker login?" ] }, { "cell_type": "markdown", "id": "2f932728-0ae0-40bf-82b7-32f7b935cf8d", "metadata": {}, "source": [ "\n", "## Training a Classification Model ##\n", "[Image Classification](https://catalog.ngc.nvidia.com/orgs/nvidia/teams/tao/models/pretrained_classification) is a popular computer vision technique in which an image is classified into one of the designated classes based on the image features. We want to train a classification model in order to label each image as either `defect` or `notdefect`. With the TAO Toolkit, we can choose the desired ResNet50 backbone as a feature extractor. As such, we will use the `classification_tf1` task, which supports the following subtasks: \n", "* `train`\n", "* `evaluate`\n", "* `inference`\n", "* `prune`\n", "* `export`\n", "\n", "

\n", " \n", "These subtasks can be invoked using the convention `tao model classification_tf1 ` on the command-line, where `args_per_subtask` are the arguments required for a given subtask. Additionally, we can always find more information about these subtasks with `tao classification_tf1 -h` or `tao classification_tf1 --help`. " ] }, { "cell_type": "markdown", "id": "281849b7-b93b-4d41-9437-8e34a2953087", "metadata": { "tags": [] }, "source": [ "\n", "### Preparation for Model Training ###\n", "For the remaining of the lab, we will use the TAO Toolkit to train a classification model. Below is what the model development workflow looks like. We start by preparing a pre-trained model and the data. Next, we prepare the configuration file(s) and begin to train the model with new data and evaluate its performance. We will export the model once its satisfactory. Note that this notebook does not include inference optimization steps, which is important for optical inspection systems that are deployed on edge devices. \n", "

" ] }, { "cell_type": "markdown", "id": "1f0bde40-85ba-472c-9984-6222fbf4d3d9", "metadata": {}, "source": [ "\n", "### Download Pre-Trained Model ###\n", "Developers typically begin by choosing and downloading a pre-trained model from [NGC](https://ngc.nvidia.com/) which contains pre-trained weights of the architecture of their choice. It's difficult to immediately identify which model/architecture will work best for a specific use case as there is often a tradeoff between time to train, accuracy, and inference performance. It is common to compare across multiple models before picking the best candidate.\n", "\n", "Here are some pointers that will help choose an appropriate model: \n", "* Look at the model inputs/outputs to decide if it will fit your use case. \n", "* Input format is also an important consideration. For example, some models expect the input to be 0-1 normalized with input channels in RGB order. Some models that use a different input order may require input preprocessing/mean subtraction that might result in suboptimal performance. \n", "\n", "We can use the `ngc registry model list ` command to get a list of models that are hosted on the NGC model registry. For example, we can use `ngc registry model list nvidia/tao/*` to list all available models. The `--column` option identifies the columns of interest. More information about the NGC Registry CLI can be found in the [User Guide](https://docs.nvidia.com/dgx/pdf/ngc-registry-cli-user-guide.pdf). The `ngc registry model download-version /[/]` command will download the model from the registry. It has a `--dest` option to specify the path to download directory. Alternatively, a catalog of support models can also be found [here](https://catalog.ngc.nvidia.com/orgs/nvidia/teams/tao/collections/tao_computervision). " ] }, { "cell_type": "code", "execution_count": 13, "id": "bc6a84c4-b633-41d1-a34d-7cd70282f348", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[{\n", " \"accuracyReached\": 77.56,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-08-18T20:15:41.146Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"153.7\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 161183816,\n", " \"versionId\": \"vgg19\"\n", "},{\n", " \"accuracyReached\": 77.17,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-08-18T20:15:13.258Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"113.2\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 118655144,\n", " \"versionId\": \"vgg16\"\n", "},{\n", " \"accuracyReached\": 65.13,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-08-18T20:18:00.942Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"6.5\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 6776712,\n", " \"versionId\": \"squeezenet\"\n", "},{\n", " \"accuracyReached\": 77.91,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-08-18T20:13:07.831Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"294.2\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 308488496,\n", " \"versionId\": \"resnet50\"\n", "},{\n", " \"accuracyReached\": 77.04,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-08-18T20:12:37.468Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"170.7\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 178944632,\n", " \"versionId\": \"resnet34\"\n", "},{\n", " \"accuracyReached\": 76.74,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-08-18T20:12:15.403Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"89.0\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 93278448,\n", " \"versionId\": \"resnet18\"\n", "},{\n", " \"accuracyReached\": 77.78,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-08-18T20:13:50.368Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"576.3\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 604328880,\n", " \"versionId\": \"resnet101\"\n", "},{\n", " \"accuracyReached\": 74.38,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-08-18T20:11:58.954Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"38.3\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 40173128,\n", " \"versionId\": \"resnet10\"\n", "},{\n", " \"accuracyReached\": 72.75,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-08-18T20:17:47.567Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"5.0\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 5258048,\n", " \"versionId\": \"mobilenet_v2\"\n", "},{\n", " \"accuracyReached\": 79.5,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-08-18T20:17:32.202Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"26.2\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 27489360,\n", " \"versionId\": \"mobilenet_v1\"\n", "},{\n", " \"accuracyReached\": 77.11,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-08-18T20:18:15.980Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"47.6\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 49958952,\n", " \"versionId\": \"googlenet\"\n", "},{\n", " \"accuracyReached\": 77.11,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-08-18T20:12:16.039Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"26.8\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 28082680,\n", " \"versionId\": \"efficientnet_b1_swish\"\n", "},{\n", " \"accuracyReached\": 77.11,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-08-18T20:11:59.218Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"26.8\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 28082608,\n", " \"versionId\": \"efficientnet_b1_relu\"\n", "},{\n", " \"accuracyReached\": 76.44,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-08-18T20:16:13.324Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"311.7\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 326824240,\n", " \"versionId\": \"darknet53\"\n", "},{\n", " \"accuracyReached\": 77.52,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-08-18T20:17:01.785Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"152.8\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 160242408,\n", " \"versionId\": \"darknet19\"\n", "},{\n", " \"accuracyReached\": 77.1,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-11-23T07:40:58.540Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"28.6\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 29955696,\n", " \"versionId\": \"cspdarknet_tiny\"\n", "},{\n", " \"accuracyReached\": 76.44,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-09-10T00:55:19.867Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"103.0\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 107993624,\n", " \"versionId\": \"cspdarknet53\"\n", "},{\n", " \"accuracyReached\": 77.52,\n", " \"batchSize\": 1,\n", " \"createdByUser\": \"n90fe0en2gvll5957fel7u75sg\",\n", " \"createdDate\": \"2021-09-10T00:55:09.123Z\",\n", " \"description\": \"\",\n", " \"gpuModel\": \"V100\",\n", " \"memoryFootprint\": \"62.9\",\n", " \"numberOfEpochs\": 80,\n", " \"status\": \"UPLOAD_COMPLETE\",\n", " \"totalFileCount\": 1,\n", " \"totalSizeInBytes\": 65913984,\n", " \"versionId\": \"cspdarknet19\"\n", "}]\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# list all available models\n", "!ngc registry model list nvidia/tao/pretrained_classification:*" ] }, { "cell_type": "markdown", "id": "9334806b-5dd7-4c36-b1a4-b2bfe2620ffe", "metadata": {}, "source": [ "

\n", "\n", "Did you get the below error message? This is likely due to a bad NGC CLI configuration. Please check the NGC CLI and Docker Registry section of the [introduction notebook](00_introduction.ipynb)." ] }, { "cell_type": "raw", "id": "3d0d9d37-d76a-420e-bfea-f1a05e902488", "metadata": {}, "source": [ "{\n", " \"error\": \"Error: Invalid apikey\"\n", "}" ] }, { "cell_type": "code", "execution_count": 14, "id": "5dd94e07-ae7f-47a1-9955-7b4415fa0a79", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "{\n", " \"download_end\": \"2026-02-02 17:30:53\",\n", " \"download_start\": \"2026-02-02 17:30:47\",\n", " \"download_time\": \"6s\",\n", " \"files_downloaded\": 1,\n", " \"local_path\": \"/dli/task/tao_project/classification/pretrained_resnet50/pretrained_classification_vresnet50\",\n", " \"size_downloaded\": \"294.2 MB\",\n", " \"status\": \"COMPLETED\"\n", "}\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# create directory to store the pre-trained model\n", "!mkdir -p $LOCAL_EXPERIMENT_DIR/pretrained_resnet50/\n", "\n", "# download the pre-trained classification model from NGC\n", "!ngc registry model download-version nvidia/tao/pretrained_classification:resnet50 --dest $LOCAL_EXPERIMENT_DIR/pretrained_resnet50 " ] }, { "cell_type": "markdown", "id": "e127c885-3369-443c-8187-03c67f426d73", "metadata": {}, "source": [ "

\n", "\n", "We designated the model to be downloaded locally to `tao_project/classification/pretrained_resnet50`, which is mapped to `/workspace/tao-experiments/classification/pretrained_resnet50` in the TAO container based on the mapping of `LOCAL_EXPERIMENT_DIR` to `TAO_EXPERIMENT_DIR`. Looking at the `local_path` and `transfer_id` keys of the output JSON, we can gather that the path of the pre-trained model should be in the `tao_project/classification/pretrained_resnet50/pretrained_classification_vresnet50` directory. When referencing paths for the TAO Toolkit, it's important to use paths based on the TAO container. In this case it would be `/workspace/tao-experiments/classification/pretrained_resnet50/pretrained_classification_vresnet50`. " ] }, { "cell_type": "code", "execution_count": 16, "id": "622a82e1-3809-41cf-8e56-3267e9459474", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "total 301272\n", "drwxr-xr-x 2 root root 4096 Feb 2 17:30 .\n", "drwxr-xr-x 3 root root 4096 Feb 2 17:30 ..\n", "-rw-r--r-- 1 root root 308488496 Feb 2 17:30 resnet_50.hdf5\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "!ls -al tao_project/classification/pretrained_resnet50/pretrained_classification_vresnet50" ] }, { "cell_type": "markdown", "id": "ab9e1652-fe1f-4a17-9307-0d2353ab1945", "metadata": { "tags": [] }, "source": [ "\n", "### Prepare Dataset ###\n", "The TAO Toolkit expects the training data for the `classification_tf1` task to be in the format described in the [documentation](https://docs.nvidia.com/tao/tao-toolkit/text/data_annotation_format.html#id8). The `classification_tf1` models expect the images to be within the directories of the respective classes. " ] }, { "cell_type": "raw", "id": "b618ccd5-3821-4550-8255-f39448a43187", "metadata": {}, "source": [ "|--dataset_root: \n", " |--train\n", " |--defect: \n", " |--DX_CX.jpg\n", " |--DX_CX.jpg\n", " |--notdefect: \n", " |--DX_CX.jpg\n", " |--DX_CX.jpg\n", " |--val\n", " |--defect: \n", " |--DX_CX.jpg\n", " |--DX_CX.jpg\n", " |--notdefect: \n", " |--DX_CX.jpg\n", " |--DX_CX.jpg\n", " |--test\n", " |--defect: \n", " |--DX_CX.jpg\n", " |--DX_CX.jpg\n", " |--notdefect: \n", " |--DX_CX.jpg\n", " |--DX_CX.jpg" ] }, { "cell_type": "markdown", "id": "f788cc67-59d7-4391-b114-e62fa44176aa", "metadata": {}, "source": [ "_The `test` folder in the above directory structure is optional; any folder can be used for inference._\n", "\n", "We start by loading the manifest of images in our dataset. " ] }, { "cell_type": "code", "execution_count": 17, "id": "5eca268c-cf0d-4517-a968-db1d13ccd571", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
true_defectdefect_img_pathdateboardcomp_idimg_shapedefect_image_namecomp_type
0notdefect/dli/task/data/AOI_DL_data_0908/0423318026324/...908423318026324C1090[54, 27, 3]D0_C1090.jpgC
1notdefect/dli/task/data/AOI_DL_data_0908/0423318026269/...908423318026269C1090[54, 27, 3]D1_C1090.jpgC
2notdefect/dli/task/data/AOI_DL_data_0908/0423318026523/...908423318026523C1090[54, 27, 3]D1_C1090.jpgC
3notdefect/dli/task/data/AOI_DL_data_0908/0423318026331/...908423318026331C1090[54, 27, 3]D1_C1090.jpgC
4notdefect/dli/task/data/AOI_DL_data_0908/0423318026211/...908423318026211C1090[53, 27, 3]D1_C1090.jpgC
\n", "
" ], "text/plain": [ " true_defect defect_img_path date \\\n", "0 notdefect /dli/task/data/AOI_DL_data_0908/0423318026324/... 908 \n", "1 notdefect /dli/task/data/AOI_DL_data_0908/0423318026269/... 908 \n", "2 notdefect /dli/task/data/AOI_DL_data_0908/0423318026523/... 908 \n", "3 notdefect /dli/task/data/AOI_DL_data_0908/0423318026331/... 908 \n", "4 notdefect /dli/task/data/AOI_DL_data_0908/0423318026211/... 908 \n", "\n", " board comp_id img_shape defect_image_name comp_type \n", "0 423318026324 C1090 [54, 27, 3] D0_C1090.jpg C \n", "1 423318026269 C1090 [54, 27, 3] D1_C1090.jpg C \n", "2 423318026523 C1090 [54, 27, 3] D1_C1090.jpg C \n", "3 423318026331 C1090 [54, 27, 3] D1_C1090.jpg C \n", "4 423318026211 C1090 [53, 27, 3] D1_C1090.jpg C " ] }, "execution_count": 17, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# export csv file as dataframe\n", "capacitor_df=pd.read_csv('capacitor_df.csv', converters={'img_shape': pd.eval})\n", "capacitor_df.head()" ] }, { "cell_type": "markdown", "id": "c536dcd8-6890-4503-879b-9409429fc98c", "metadata": {}, "source": [ "Below we will split the data into `train` set and `validation` set and copy the images into their respective folder. " ] }, { "cell_type": "code", "execution_count": 18, "id": "96b5ee50-6c69-4b9a-95bd-980323feaa4b", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "/tmp/ipykernel_3011/4020387363.py:6: DeprecationWarning: DataFrameGroupBy.apply operated on the grouping columns. This behavior is deprecated, and in a future version of pandas the grouping columns will be excluded from the operation. Either pass `include_groups=False` to exclude the groupings or explicitly select the grouping columns after groupby to silence this warning.\n", " val_set=capacitor_df.groupby('true_defect', group_keys=False).apply(lambda x: x.sample(frac=0.3))\n" ] }, { "data": { "text/plain": [ "dataset true_defect\n", "train defect 69\n", " notdefect 1263\n", "val defect 30\n", " notdefect 541\n", "dtype: int64" ] }, "execution_count": 18, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# set default as training set\n", "capacitor_df['dataset']='train'\n", "\n", "# sample 30% and set as validation set\n", "val_set=capacitor_df.groupby('true_defect', group_keys=False).apply(lambda x: x.sample(frac=0.3))\n", "capacitor_df.loc[val_set.index, 'dataset']='val'\n", "capacitor_df.groupby(['dataset', 'true_defect']).size()" ] }, { "cell_type": "code", "execution_count": 19, "id": "9e959942-b69c-41aa-accc-1443b10b5e08", "metadata": {}, "outputs": [], "source": [ "# DO NOT CHANGE THIS CELL\n", "# remove existing data from previous experiment (if any)\n", "!rm -rf $LOCAL_DATA_DIR/*\n", "\n", "!mkdir -p $LOCAL_DATA_DIR/train/defect\n", "!mkdir -p $LOCAL_DATA_DIR/train/notdefect\n", "!mkdir -p $LOCAL_DATA_DIR/val/defect\n", "!mkdir -p $LOCAL_DATA_DIR/val/notdefect" ] }, { "cell_type": "code", "execution_count": 20, "id": "fd4be461-0ec1-4b50-8d68-912840f42f8e", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "It took 1.9 seconds to copy 1903 images.\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# time the process\n", "start=time.time()\n", "\n", "# iterate through the capacitors\n", "for idx, row in capacitor_df.iterrows(): \n", " shutil.copyfile(row['defect_img_path'], f\"{os.environ['LOCAL_DATA_DIR']}/{row['dataset']}/{row['true_defect']}/{row['date']}_{row['board']}_{row['defect_image_name']}\")\n", "\n", "print('It took {} seconds to copy {} images.'.format(round(time.time()-start, 2), len(capacitor_df)))" ] }, { "cell_type": "markdown", "id": "a1a34fb6-3cb3-41a0-be9f-6cc09dc74437", "metadata": {}, "source": [ "Let's take a quick look at the size of the images. " ] }, { "cell_type": "code", "execution_count": 21, "id": "121ce22c-6a6b-4651-9193-77a3a7ff2a81", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "The median dimension is 56.0 x 27.0.\n" ] }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# plot image dimensions\n", "plt.figure(figsize=(5, 5))\n", "plt.scatter(capacitor_df['img_shape'].str[0], \n", " capacitor_df['img_shape'].str[1])\n", "\n", "print('The median dimension is {} x {}.'.format(np.median(capacitor_df['img_shape'].str[0]), np.median(capacitor_df['img_shape'].str[1])))" ] }, { "cell_type": "markdown", "id": "d21b8bd7-490e-4ee7-a3ed-3930e474cbbb", "metadata": {}, "source": [ "The images range in size and dimension. However, the TAO Toolkit will resize the images in the dataset to a specific shape when fed to the model for training. " ] }, { "cell_type": "markdown", "id": "a7b710aa-c92c-4d09-934f-5096b2711965", "metadata": {}, "source": [ "\n", "### Model Training ###\n", "Training configuration is done through an experiment spec file, which includes options such as which dataset to use for training, which dataset to use for validation, which pre-trained model architecture to use, which hyperparameters to tune, and other training options. The `train`, `evaluate`, `prune`, and `inference` subtasks for a `classification_tf1` experiment share the same configuration file. Configuration files can be created from scratch or modified using the templates provided in TAO Toolkit's [sample applications](https://docs.nvidia.com/tao/tao-toolkit/#cv-applications). \n", "\n", "The training configuration file has three sections: \n", "* `model_config` (required for training)\n", "* `eval_config` (not required for training)\n", "* `training_config` (required for training)" ] }, { "cell_type": "markdown", "id": "3fb34518-d2fd-449c-bcbf-fa5b2fcd086f", "metadata": {}, "source": [ "

\n", "\n", "We will create the configuration files using templates. Specifically, we have broken the configuration files into separate parts for ease of discussion, which we will combine at the end for the TAO Toolkit to consume. \n", "\n", "Execute the below cells to preview the combined training/evaluation configuration file that will be used. It is currently not usable as we have made some intentional modifications that will require correction. Furthermore, the `eval_config` portion of the combined configuration file is not used for training, so leaving it unmodified will not adversely impact the training process. " ] }, { "cell_type": "code", "execution_count": 22, "id": "b3a3201c-2443-4f94-852f-f15cc51b25e1", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "model_config {\n", " arch: \"<<<>>>\"\n", " n_layers: <<<>>>\n", " freeze_blocks: 0\n", " use_batch_norm: True\n", " input_image_size: \"3,224,224\"\n", "}\n", "##################################################\n", "train_config {\n", " train_dataset_path: \"/workspace/tao-experiments/data/<<<>>>\"\n", " val_dataset_path: \"/workspace/tao-experiments/data/<<<>>>\"\n", " pretrained_model_path: \"/workspace/tao-experiments/classification/pretrained_resnet50/pretrained_classification_vresnet50/<<<>>>.hdf5\"\n", " optimizer {\n", " sgd {\n", " lr: 0.01\n", " decay: 0.0\n", " momentum: 0.9\n", " nesterov: False\n", " }\n", " }\n", " n_epochs: <<<>>>\n", " batch_size_per_gpu: 32\n", " n_workers: 8\n", " enable_random_crop: False\n", " enable_center_crop: False\n", " enable_color_augmentation: False\n", " preprocess_mode: \"caffe\"\n", " reg_config {\n", " type: \"L2\"\n", " scope: \"Conv2D, Dense\"\n", " weight_decay: 0.00005\n", " }\n", " lr_config {\n", " step {\n", " learning_rate: 0.006\n", " step_size: 10\n", " gamma: 0.1\n", " }\n", " }\n", "}\n", "##################################################\n", "eval_config {\n", " eval_dataset_path: \"/workspace/tao-experiments/data/<<<>>>\"\n", " model_path: \"/workspace/tao-experiments/classification/resnet50/weights/<<<>>>.hdf5\"\n", " top_k: 1\n", " batch_size: 32\n", " n_workers: 8\n", " enable_center_crop: False\n", "}\n", "##################################################\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# combining configuration components in separate files and writing into one\n", "!cat $LOCAL_SPECS_DIR/resnet50/model_config.txt \\\n", " $LOCAL_SPECS_DIR/resnet50/train_config.txt \\\n", " $LOCAL_SPECS_DIR/resnet50/eval_config.txt \\\n", " > $LOCAL_SPECS_DIR/resnet50/combined_config.txt\n", "!cat $LOCAL_SPECS_DIR/resnet50/combined_config.txt" ] }, { "cell_type": "markdown", "id": "02ded461-e7f3-4d20-b406-e1a0b309eb29", "metadata": {}, "source": [ "

\n", "\n", "Note that we must leave an empty new line at the end of each text file to ensure the `combined_config.txt` is created properly. " ] }, { "cell_type": "markdown", "id": "a09b587b-d099-48ab-be87-3e3c8b3f41af", "metadata": { "tags": [] }, "source": [ "\n", "### Exercise #2 - Modify Model Config ###\n", "The classification model can be configured using the `model_config` option in the spec file. \n", "* `arch (str)`: This defines the architecture of the back bone feature extractor to be used to train.\n", "* `n_layers (int)`: Depth of the feature extractor for scalable templates. \n", "* `use_batch_norm (bool)`: Boolean variable to use batch normalization layers or not. \n", "* `freeze_blocks (float repeated)`: This parameter defines which blocks may be frozen from the instantiated feature extractor template, and is different for different feature extractor templates.\n", "* `input_image_size (str)`: The dimension of the input layer of the model. Images in the dataset will be resized to this shape by the dataloader when fed to the model for training. \n", "\n", "**Instructions**:
\n", "* Modify the `model_config`[(here)](tao_project/spec_files/resnet50/model_config.txt) section of the training configuration file by changing the ``s into acceptable values and **save changes**. \n", " * For this lab, we will use _ResNet50_ as the architecture for the classification model. [Residual neural network](https://en.wikipedia.org/wiki/Residual_neural_network), or **ResNet**, is a type of convolutional neural network used as a backbone for many computer vision tasks. The `50` refers to the number of layers in this architecture. It should be noted that typically the deeper (i.e. more layers) a neural network is, the more time consuming it is to compute. " ] }, { "cell_type": "code", "execution_count": 23, "id": "251fa570-af61-4f08-a658-ca8caca46540", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "model_config {\n", " arch: \"resnet\"\n", " n_layers: 50\n", " freeze_blocks: 0\n", " use_batch_norm: True\n", " input_image_size: \"3,224,224\"\n", "}" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# read the config file\n", "!cat $LOCAL_SPECS_DIR/resnet50/model_config.txt" ] }, { "cell_type": "markdown", "id": "da9eb38b-7041-4078-b4ad-60ea7086e2fd", "metadata": { "tags": [] }, "source": [ "\n", "### Exercise #3 - Modify Train Config ###\n", "The `training_config` describes the training and learning process. \n", "* `train_dataset_path (str)`: UNIX format path to the root directory of the training dataset.\n", "* `val_dataset_path (str)`: UNIX format path to the root directory of the validation dataset.\n", "* `pretrained_model_path (str)`: UNIX format path to the model file containing the pretrained weights to initialize the model from.\n", " * Based on how NGC names the model downloaded, we should use `/workspace/tao-experiments/classification/pretrained_resnet50/pretrained_classification_vresnet50/resnet_50.hdf5` to reference the pre-trained model. \n", "* `optimizer (dict)`: This parameter defines which optimizer to use for training. Can be chosen from `sgd`, `adam`, or `rmsprop`.\n", " * `sgd (dict)`\n", " * `lr (float)`\n", " * `decay (float)`\n", " * `momentum (float)`\n", " * `nesterov (bool)`\n", "* `batch_size_per_gpu (int)`: This parameter defines the number of images per batch per gpu.\n", "* `n_workers (int)`: Number of workers fetching batches of images in the training/validation dataloader.\n", "* `n_epochs (int)`: This parameter defines the total number of epochs to run the experiment.\n", "* `enable_random_crop (bool)`: A flag to enable random crop during training.\n", "* `enable_center_crop (bool)`: A flag to enable center crop during validation.\n", "* `enable_color_augmentation (bool)`: A flag to enable color augmentation during training.\n", "* `preprocess_mode (str)`: Mode for input image preprocessing. Defaults to `caffe`.\n", "* `reg_config (dict)`: The parameters for regularizers.\n", " * `type (str)`: Regularizer type can be `L1`, `L2` or `None`.\n", " * `scope (str)`: Scope can be either `Conv2D` or `Dense` or both.\n", " * `weight_decay (float)`: 0 < weight decay < 1. \n", "* `lr_config (dict)`: The parameters for learning rate scheduler.\n", " * `step (dict)`: Defines the step learning rate scheduler. \n", " * `learning_rate (float)`: The base (maximum) learning rate value.\n", " * `step_size (int)`: The progress (percentage of the entire training duration) after which the learning rate will be decreased.\n", " * `gamma (float)`: The multiplicative factor used to decrease the learning rate.\n", "\n", "**Instructions**:
\n", "* Modify the `train_config`[(here)](tao_project/spec_files/resnet50/train_config.txt) section of the configuration file by changing the ``s into acceptable values and **save changes**. Typically, using a higher epochs count will improve model performance but takes longer time to complete. For the purpose of this exercise, we recommend starting with a low `n_epochs`, such as `10`, to allow the model to converge without taking too much time. " ] }, { "cell_type": "code", "execution_count": 24, "id": "d5ac8ebc-0ede-4d6f-b982-47f6bf649b95", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "train_config {\n", " train_dataset_path: \"/workspace/tao-experiments/data/train\"\n", " val_dataset_path: \"/workspace/tao-experiments/data/val\"\n", " pretrained_model_path: \"/workspace/tao-experiments/classification/pretrained_resnet50/pretrained_classification_vresnet50/resnet_50.hdf5\"\n", " optimizer {\n", " sgd {\n", " lr: 0.01\n", " decay: 0.0\n", " momentum: 0.9\n", " nesterov: False\n", " }\n", " }\n", " n_epochs: 10\n", " batch_size_per_gpu: 32\n", " n_workers: 8\n", " enable_random_crop: False\n", " enable_center_crop: False\n", " enable_color_augmentation: False\n", " preprocess_mode: \"caffe\"\n", " reg_config {\n", " type: \"L2\"\n", " scope: \"Conv2D, Dense\"\n", " weight_decay: 0.00005\n", " }\n", " lr_config {\n", " step {\n", " learning_rate: 0.006\n", " step_size: 10\n", " gamma: 0.1\n", " }\n", " }\n", "}\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# read the config file\n", "!cat $LOCAL_SPECS_DIR/resnet50/train_config.txt" ] }, { "cell_type": "markdown", "id": "2cac10ab-9555-42fa-bf53-8a4c81daaedc", "metadata": {}, "source": [ "\n", "### Combine Configuration Files and Initiate Model Training ###" ] }, { "cell_type": "code", "execution_count": 25, "id": "67af8d63-2748-47b9-9bc1-ac561a34d6af", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "model_config {\n", " arch: \"resnet\"\n", " n_layers: 50\n", " freeze_blocks: 0\n", " use_batch_norm: True\n", " input_image_size: \"3,224,224\"\n", "}train_config {\n", " train_dataset_path: \"/workspace/tao-experiments/data/train\"\n", " val_dataset_path: \"/workspace/tao-experiments/data/val\"\n", " pretrained_model_path: \"/workspace/tao-experiments/classification/pretrained_resnet50/pretrained_classification_vresnet50/resnet_50.hdf5\"\n", " optimizer {\n", " sgd {\n", " lr: 0.01\n", " decay: 0.0\n", " momentum: 0.9\n", " nesterov: False\n", " }\n", " }\n", " n_epochs: 10\n", " batch_size_per_gpu: 32\n", " n_workers: 8\n", " enable_random_crop: False\n", " enable_center_crop: False\n", " enable_color_augmentation: False\n", " preprocess_mode: \"caffe\"\n", " reg_config {\n", " type: \"L2\"\n", " scope: \"Conv2D, Dense\"\n", " weight_decay: 0.00005\n", " }\n", " lr_config {\n", " step {\n", " learning_rate: 0.006\n", " step_size: 10\n", " gamma: 0.1\n", " }\n", " }\n", "}\n", "eval_config {\n", " eval_dataset_path: \"/workspace/tao-experiments/data/<<<>>>\"\n", " model_path: \"/workspace/tao-experiments/classification/resnet50/weights/<<<>>>.hdf5\"\n", " top_k: 1\n", " batch_size: 32\n", " n_workers: 8\n", " enable_center_crop: False\n", "}\n", "##################################################\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# combining configuration components in separate files and writing into one\n", "!cat $LOCAL_SPECS_DIR/resnet50/model_config.txt \\\n", " $LOCAL_SPECS_DIR/resnet50/train_config.txt \\\n", " $LOCAL_SPECS_DIR/resnet50/eval_config.txt \\\n", " > $LOCAL_SPECS_DIR/resnet50/combined_config.txt\n", "!cat $LOCAL_SPECS_DIR/resnet50/combined_config.txt" ] }, { "cell_type": "markdown", "id": "80fcd57b-41cf-432e-b2f3-033140bb8799", "metadata": {}, "source": [ "After preparing the input data and setting up the spec file(s), we can start training the classification model." ] }, { "cell_type": "raw", "id": "71e40905-6482-48dc-b81b-a5a5a9657f99", "metadata": {}, "source": [ "tao model classification train [-h] -e \n", " -r \n", " -k \n", " [-v Set Verbosity of the logger]\n", " [--gpus GPUS]\n", " [--gpu_index GPU_INDEX]" ] }, { "cell_type": "markdown", "id": "fbbe0274-1990-48cf-b3d1-191d9e191f67", "metadata": {}, "source": [ "When using the `train` subtask, the `-e` argument indicates the path to the spec file, the `-r` argument indicates the result directory, and the `-k` indicates the key to _load_ the pre-trained weights. \n", "\n", "_Multi-GPU support can be enabled for those with the hardware using the `--gpus` argument. When running the training with more than one GPU, we will need to modify the `batch_size` and `learning_rate`. In most cases, scaling down the batch-size by a factor of NUM_GPU's or scaling up the learning rate by a factor of NUM_GPUs would be a good place to start._" ] }, { "cell_type": "code", "execution_count": 26, "id": "890df497-2679-4f97-a175-7d76ad520d7c", "metadata": {}, "outputs": [], "source": [ "# DO NOT CHANGE THIS CELL\n", "# remove any previous training if exists\n", "!rm -rf $LOCAL_EXPERIMENT_DIR/resnet50" ] }, { "cell_type": "markdown", "id": "fab8ba85-1a06-424e-a4ec-6461de7df0d2", "metadata": {}, "source": [ "

\n", "\n", "While the TAO Toolkit is running, there may be some _TensorFlow deprecation_ warning messages that can be ignored. " ] }, { "cell_type": "code", "execution_count": 27, "id": "23671788-5215-4ab0-b6b4-09d9ab6738cd", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "2026-02-02 17:51:08,008 [TAO Toolkit] [INFO] root 160: Registry: ['nvcr.io']\n", "2026-02-02 17:51:08,104 [TAO Toolkit] [INFO] nvidia_tao_cli.components.instance_handler.local_instance 360: Running command in container: nvcr.io/nvidia/tao/tao-toolkit:5.0.0-tf1.15.5\n", "2026-02-02 17:51:08,116 [TAO Toolkit] [INFO] nvidia_tao_cli.components.docker_handler.docker_handler 301: Printing tty value True\n", "Using TensorFlow backend.\n", "2026-02-02 17:51:08.922693: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 17:51:08,973 [TAO Toolkit] [WARNING] tensorflow 40: Deprecation warnings have been disabled. Set TF_ENABLE_DEPRECATION_WARNINGS=1 to re-enable them.\n", "2026-02-02 17:51:10,031 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use sklearn by default. This improves performance in some cases. To enable sklearn export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 17:51:10,085 [TAO Toolkit] [WARNING] tensorflow 42: TensorFlow will not use Dask by default. This improves performance in some cases. To enable Dask export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 17:51:10,090 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use Pandas by default. This improves performance in some cases. To enable Pandas export the environment variable TF_ALLOW_IOLIBS=1.\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:150: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_hue(img, max_delta=10.0):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:173: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_saturation(img, max_shift):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:183: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_contrast(img, center, max_contrast_scale):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:192: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_shift(x_img, shift_stddev):\n", "2026-02-02 17:51:11.914350: I tensorflow/core/platform/profile_utils/cpu_utils.cc:109] CPU Frequency: 2499995000 Hz\n", "2026-02-02 17:51:11.914764: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x8eac0f0 initialized for platform Host (this does not guarantee that XLA will be used). Devices:\n", "2026-02-02 17:51:11.914803: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version\n", "2026-02-02 17:51:11.916154: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcuda.so.1\n", "2026-02-02 17:51:12.101379: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:51:12.103285: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x8cc74a0 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:\n", "2026-02-02 17:51:12.103326: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Tesla T4, Compute Capability 7.5\n", "2026-02-02 17:51:12.103671: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:51:12.105440: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1674] Found device 0 with properties: \n", "name: Tesla T4 major: 7 minor: 5 memoryClockRate(GHz): 1.59\n", "pciBusID: 0000:00:1e.0\n", "2026-02-02 17:51:12.105508: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 17:51:12.105642: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcublas.so.12\n", "2026-02-02 17:51:12.107815: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcufft.so.11\n", "2026-02-02 17:51:12.107931: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcurand.so.10\n", "2026-02-02 17:51:12.111960: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcusolver.so.11\n", "2026-02-02 17:51:12.113208: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcusparse.so.12\n", "2026-02-02 17:51:12.113286: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudnn.so.8\n", "2026-02-02 17:51:12.113417: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:51:12.115212: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:51:12.116912: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1802] Adding visible gpu devices: 0\n", "2026-02-02 17:51:12.116958: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 17:51:12.125622: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1214] Device interconnect StreamExecutor with strength 1 edge matrix:\n", "2026-02-02 17:51:12.125651: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1220] 0 \n", "2026-02-02 17:51:12.125665: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1233] 0: N \n", "2026-02-02 17:51:12.125846: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:51:12.127636: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:51:12.129384: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1359] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 13496 MB memory) -> physical GPU (device: 0, name: Tesla T4, pci bus id: 0000:00:1e.0, compute capability: 7.5)\n", "Using TensorFlow backend.\n", "WARNING:tensorflow:Deprecation warnings have been disabled. Set TF_ENABLE_DEPRECATION_WARNINGS=1 to re-enable them.\n", "WARNING:tensorflow:TensorFlow will not use sklearn by default. This improves performance in some cases. To enable sklearn export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 17:51:13,795 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use sklearn by default. This improves performance in some cases. To enable sklearn export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING:tensorflow:TensorFlow will not use Dask by default. This improves performance in some cases. To enable Dask export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 17:51:13,833 [TAO Toolkit] [WARNING] tensorflow 42: TensorFlow will not use Dask by default. This improves performance in some cases. To enable Dask export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING:tensorflow:TensorFlow will not use Pandas by default. This improves performance in some cases. To enable Pandas export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 17:51:13,837 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use Pandas by default. This improves performance in some cases. To enable Pandas export the environment variable TF_ALLOW_IOLIBS=1.\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:150: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_hue(img, max_delta=10.0):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:173: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_saturation(img, max_shift):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:183: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_contrast(img, center, max_contrast_scale):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:192: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_shift(x_img, shift_stddev):\n", "2026-02-02 17:51:15,055 [TAO Toolkit] [INFO] __main__ 388: Loading experiment spec at /workspace/tao-experiments/spec_files/resnet50/combined_config.txt.\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/train.py:398: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n", "\n", "2026-02-02 17:51:15,059 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/train.py:398: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/train.py:407: The name tf.Session is deprecated. Please use tf.compat.v1.Session instead.\n", "\n", "2026-02-02 17:51:15,059 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/train.py:407: The name tf.Session is deprecated. Please use tf.compat.v1.Session instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/train.py:431: The name tf.logging.set_verbosity is deprecated. Please use tf.compat.v1.logging.set_verbosity instead.\n", "\n", "2026-02-02 17:51:15,404 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/train.py:431: The name tf.logging.set_verbosity is deprecated. Please use tf.compat.v1.logging.set_verbosity instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/train.py:431: The name tf.logging.INFO is deprecated. Please use tf.compat.v1.logging.INFO instead.\n", "\n", "2026-02-02 17:51:15,404 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/train.py:431: The name tf.logging.INFO is deprecated. Please use tf.compat.v1.logging.INFO instead.\n", "\n", "2026-02-02 17:51:15,405 [TAO Toolkit] [INFO] __main__ 478: Default image mean [103.939, 116.779, 123.68] will be used.\n", "Found 1332 images belonging to 2 classes.\n", "2026-02-02 17:51:15,433 [TAO Toolkit] [INFO] __main__ 294: Processing dataset (train): /workspace/tao-experiments/data/train\n", "Found 571 images belonging to 2 classes.\n", "2026-02-02 17:51:15,442 [TAO Toolkit] [INFO] __main__ 311: Processing dataset (validation): /workspace/tao-experiments/data/val\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n", "\n", "2026-02-02 17:51:15,442 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n", "\n", "2026-02-02 17:51:15,443 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n", "\n", "2026-02-02 17:51:15,444 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:1834: The name tf.nn.fused_batch_norm is deprecated. Please use tf.compat.v1.nn.fused_batch_norm instead.\n", "\n", "2026-02-02 17:51:15,460 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:1834: The name tf.nn.fused_batch_norm is deprecated. Please use tf.compat.v1.nn.fused_batch_norm instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n", "\n", "2026-02-02 17:51:15,465 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/third_party/keras/tensorflow_backend.py:199: The name tf.nn.avg_pool is deprecated. Please use tf.nn.avg_pool2d instead.\n", "\n", "2026-02-02 17:51:17,000 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/third_party/keras/tensorflow_backend.py:199: The name tf.nn.avg_pool is deprecated. Please use tf.nn.avg_pool2d instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:174: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n", "\n", "2026-02-02 17:51:18,690 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:174: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:190: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n", "\n", "2026-02-02 17:51:18,690 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:190: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:199: The name tf.is_variable_initialized is deprecated. Please use tf.compat.v1.is_variable_initialized instead.\n", "\n", "2026-02-02 17:51:18,690 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:199: The name tf.is_variable_initialized is deprecated. Please use tf.compat.v1.is_variable_initialized instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:206: The name tf.variables_initializer is deprecated. Please use tf.compat.v1.variables_initializer instead.\n", "\n", "2026-02-02 17:51:19,742 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:206: The name tf.variables_initializer is deprecated. Please use tf.compat.v1.variables_initializer instead.\n", "\n", "__________________________________________________________________________________________________\n", "Layer (type) Output Shape Param # Connected to \n", "==================================================================================================\n", "input_1 (InputLayer) (None, 3, 224, 224) 0 \n", "__________________________________________________________________________________________________\n", "conv1 (Conv2D) (None, 64, 112, 112) 9408 input_1[0][0] \n", "__________________________________________________________________________________________________\n", "bn_conv1 (BatchNormalization) (None, 64, 112, 112) 256 conv1[0][0] \n", "__________________________________________________________________________________________________\n", "activation_1 (Activation) (None, 64, 112, 112) 0 bn_conv1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_1 (Conv2D) (None, 64, 56, 56) 4096 activation_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu_1 (Activation) (None, 64, 56, 56) 0 block_1a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu_2 (Activation) (None, 64, 56, 56) 0 block_1a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_shortcut (Conv2D) (None, 256, 56, 56) 16384 activation_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_shortcut (BatchNorm (None, 256, 56, 56) 1024 block_1a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_1 (Add) (None, 256, 56, 56) 0 block_1a_bn_3[0][0] \n", " block_1a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu (Activation) (None, 256, 56, 56) 0 add_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_1 (Conv2D) (None, 64, 56, 56) 16384 block_1a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu_1 (Activation) (None, 64, 56, 56) 0 block_1b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu_2 (Activation) (None, 64, 56, 56) 0 block_1b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_2 (Add) (None, 256, 56, 56) 0 block_1b_bn_3[0][0] \n", " block_1a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu (Activation) (None, 256, 56, 56) 0 add_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_1 (Conv2D) (None, 64, 56, 56) 16384 block_1b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu_1 (Activation) (None, 64, 56, 56) 0 block_1c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu_2 (Activation) (None, 64, 56, 56) 0 block_1c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_3 (Add) (None, 256, 56, 56) 0 block_1c_bn_3[0][0] \n", " block_1b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu (Activation) (None, 256, 56, 56) 0 add_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_1 (Conv2D) (None, 128, 28, 28) 32768 block_1c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu_1 (Activation) (None, 128, 28, 28) 0 block_2a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu_2 (Activation) (None, 128, 28, 28) 0 block_2a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_shortcut (Conv2D) (None, 512, 28, 28) 131072 block_1c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_shortcut (BatchNorm (None, 512, 28, 28) 2048 block_2a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_4 (Add) (None, 512, 28, 28) 0 block_2a_bn_3[0][0] \n", " block_2a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu (Activation) (None, 512, 28, 28) 0 add_4[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu_1 (Activation) (None, 128, 28, 28) 0 block_2b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu_2 (Activation) (None, 128, 28, 28) 0 block_2b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_5 (Add) (None, 512, 28, 28) 0 block_2b_bn_3[0][0] \n", " block_2a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu (Activation) (None, 512, 28, 28) 0 add_5[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu_1 (Activation) (None, 128, 28, 28) 0 block_2c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu_2 (Activation) (None, 128, 28, 28) 0 block_2c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_6 (Add) (None, 512, 28, 28) 0 block_2c_bn_3[0][0] \n", " block_2b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu (Activation) (None, 512, 28, 28) 0 add_6[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2d_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu_1 (Activation) (None, 128, 28, 28) 0 block_2d_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2d_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2d_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu_2 (Activation) (None, 128, 28, 28) 0 block_2d_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2d_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2d_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_7 (Add) (None, 512, 28, 28) 0 block_2d_bn_3[0][0] \n", " block_2c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu (Activation) (None, 512, 28, 28) 0 add_7[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_1 (Conv2D) (None, 256, 14, 14) 131072 block_2d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu_1 (Activation) (None, 256, 14, 14) 0 block_3a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu_2 (Activation) (None, 256, 14, 14) 0 block_3a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_shortcut (Conv2D) (None, 1024, 14, 14) 524288 block_2d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_shortcut (BatchNorm (None, 1024, 14, 14) 4096 block_3a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_8 (Add) (None, 1024, 14, 14) 0 block_3a_bn_3[0][0] \n", " block_3a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu (Activation) (None, 1024, 14, 14) 0 add_8[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu_1 (Activation) (None, 256, 14, 14) 0 block_3b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu_2 (Activation) (None, 256, 14, 14) 0 block_3b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_9 (Add) (None, 1024, 14, 14) 0 block_3b_bn_3[0][0] \n", " block_3a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu (Activation) (None, 1024, 14, 14) 0 add_9[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu_1 (Activation) (None, 256, 14, 14) 0 block_3c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu_2 (Activation) (None, 256, 14, 14) 0 block_3c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_10 (Add) (None, 1024, 14, 14) 0 block_3c_bn_3[0][0] \n", " block_3b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu (Activation) (None, 1024, 14, 14) 0 add_10[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3d_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu_1 (Activation) (None, 256, 14, 14) 0 block_3d_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3d_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3d_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu_2 (Activation) (None, 256, 14, 14) 0 block_3d_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3d_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3d_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_11 (Add) (None, 1024, 14, 14) 0 block_3d_bn_3[0][0] \n", " block_3c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu (Activation) (None, 1024, 14, 14) 0 add_11[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3e_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu_1 (Activation) (None, 256, 14, 14) 0 block_3e_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3e_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3e_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu_2 (Activation) (None, 256, 14, 14) 0 block_3e_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3e_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3e_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_12 (Add) (None, 1024, 14, 14) 0 block_3e_bn_3[0][0] \n", " block_3d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu (Activation) (None, 1024, 14, 14) 0 add_12[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3e_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3f_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu_1 (Activation) (None, 256, 14, 14) 0 block_3f_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3f_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3f_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu_2 (Activation) (None, 256, 14, 14) 0 block_3f_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3f_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3f_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_13 (Add) (None, 1024, 14, 14) 0 block_3f_bn_3[0][0] \n", " block_3e_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu (Activation) (None, 1024, 14, 14) 0 add_13[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_1 (Conv2D) (None, 512, 14, 14) 524288 block_3f_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu_1 (Activation) (None, 512, 14, 14) 0 block_4a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu_2 (Activation) (None, 512, 14, 14) 0 block_4a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_shortcut (Conv2D) (None, 2048, 14, 14) 2097152 block_3f_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_shortcut (BatchNorm (None, 2048, 14, 14) 8192 block_4a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_14 (Add) (None, 2048, 14, 14) 0 block_4a_bn_3[0][0] \n", " block_4a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu (Activation) (None, 2048, 14, 14) 0 add_14[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_1 (Conv2D) (None, 512, 14, 14) 1048576 block_4a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu_1 (Activation) (None, 512, 14, 14) 0 block_4b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu_2 (Activation) (None, 512, 14, 14) 0 block_4b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_15 (Add) (None, 2048, 14, 14) 0 block_4b_bn_3[0][0] \n", " block_4a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu (Activation) (None, 2048, 14, 14) 0 add_15[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_1 (Conv2D) (None, 512, 14, 14) 1048576 block_4b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu_1 (Activation) (None, 512, 14, 14) 0 block_4c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu_2 (Activation) (None, 512, 14, 14) 0 block_4c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_16 (Add) (None, 2048, 14, 14) 0 block_4c_bn_3[0][0] \n", " block_4b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu (Activation) (None, 2048, 14, 14) 0 add_16[0][0] \n", "__________________________________________________________________________________________________\n", "avg_pool (AveragePooling2D) (None, 2048, 1, 1) 0 block_4c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "flatten (Flatten) (None, 2048) 0 avg_pool[0][0] \n", "__________________________________________________________________________________________________\n", "predictions (Dense) (None, 2) 4098 flatten[0][0] \n", "==================================================================================================\n", "Total params: 23,565,250\n", "Trainable params: 23,502,722\n", "Non-trainable params: 62,528\n", "__________________________________________________________________________________________________\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n", "\n", "2026-02-02 17:52:27,073 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n", "\n", "2026-02-02 17:52:27,084 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/common/utils.py:1133: The name tf.summary.FileWriter is deprecated. Please use tf.compat.v1.summary.FileWriter instead.\n", "\n", "2026-02-02 17:52:27,114 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/common/utils.py:1133: The name tf.summary.FileWriter is deprecated. Please use tf.compat.v1.summary.FileWriter instead.\n", "\n", "2026-02-02 17:52:27,115 [TAO Toolkit] [INFO] nvidia_tao_tf1.cv.common.logging.logging 197: Log file already exists at /workspace/tao-experiments/classification/resnet50/status.json\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:986: The name tf.assign_add is deprecated. Please use tf.compat.v1.assign_add instead.\n", "\n", "2026-02-02 17:52:30,267 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:986: The name tf.assign_add is deprecated. Please use tf.compat.v1.assign_add instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:973: The name tf.assign is deprecated. Please use tf.compat.v1.assign instead.\n", "\n", "2026-02-02 17:52:30,589 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:973: The name tf.assign is deprecated. Please use tf.compat.v1.assign instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/common/utils.py:1181: The name tf.summary.merge_all is deprecated. Please use tf.compat.v1.summary.merge_all instead.\n", "\n", "2026-02-02 17:52:33,114 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/common/utils.py:1181: The name tf.summary.merge_all is deprecated. Please use tf.compat.v1.summary.merge_all instead.\n", "\n", "2026-02-02 17:52:34,924 [TAO Toolkit] [INFO] root 2102: Starting Training Loop.\n", "Epoch 1/10\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/common/utils.py:199: The name tf.Summary is deprecated. Please use tf.compat.v1.Summary instead.\n", "\n", "2026-02-02 17:53:01,589 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/common/utils.py:199: The name tf.Summary is deprecated. Please use tf.compat.v1.Summary instead.\n", "\n", "42/42 [==============================] - 59s 1s/step - loss: 0.4340 - acc: 0.9274 - val_loss: 0.3599 - val_acc: 0.9475\n", "2026-02-02 17:53:48,745 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 2/10\n", "42/42 [==============================] - 19s 448ms/step - loss: 0.3400 - acc: 0.9482 - val_loss: 0.3454 - val_acc: 0.9475\n", "2026-02-02 17:54:14,829 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 3/10\n", "42/42 [==============================] - 19s 447ms/step - loss: 0.3342 - acc: 0.9486 - val_loss: 0.3448 - val_acc: 0.9475\n", "2026-02-02 17:54:35,050 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 4/10\n", "42/42 [==============================] - 19s 446ms/step - loss: 0.3388 - acc: 0.9478 - val_loss: 0.3440 - val_acc: 0.9475\n", "2026-02-02 17:54:54,259 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 5/10\n", "42/42 [==============================] - 19s 447ms/step - loss: 0.3325 - acc: 0.9478 - val_loss: 0.3447 - val_acc: 0.9475\n", "2026-02-02 17:55:13,509 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 6/10\n", "42/42 [==============================] - 19s 448ms/step - loss: 0.3342 - acc: 0.9486 - val_loss: 0.3441 - val_acc: 0.9475\n", "2026-02-02 17:55:32,768 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 7/10\n", "42/42 [==============================] - 19s 448ms/step - loss: 0.3328 - acc: 0.9482 - val_loss: 0.3439 - val_acc: 0.9475\n", "2026-02-02 17:55:52,032 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 8/10\n", "42/42 [==============================] - 19s 448ms/step - loss: 0.3376 - acc: 0.9486 - val_loss: 0.3449 - val_acc: 0.9475\n", "2026-02-02 17:56:11,359 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 9/10\n", "42/42 [==============================] - 19s 448ms/step - loss: 0.3358 - acc: 0.9482 - val_loss: 0.3451 - val_acc: 0.9475\n", "2026-02-02 17:56:30,639 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 10/10\n", "42/42 [==============================] - 19s 449ms/step - loss: 0.3380 - acc: 0.9478 - val_loss: 0.3446 - val_acc: 0.9475\n", "2026-02-02 17:56:49,950 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "2026-02-02 17:56:49,966 [TAO Toolkit] [INFO] root 2102: Training loop complete.\n", "2026-02-02 17:56:49,978 [TAO Toolkit] [INFO] root 2102: Final model evaluation in progress.\n", "2026-02-02 17:56:53,749 [TAO Toolkit] [INFO] root 2102: Model evaluation is complete.\n", "2026-02-02 17:56:53,749 [TAO Toolkit] [INFO] __main__ 625: Total Val Loss: 0.34459400177001953\n", "2026-02-02 17:56:53,749 [TAO Toolkit] [INFO] __main__ 626: Total Val accuracy: 0.9474605917930603\n", "2026-02-02 17:56:53,749 [TAO Toolkit] [INFO] root 2102: Training finished successfully.\n", "2026-02-02 17:56:53,749 [TAO Toolkit] [INFO] __main__ 651: Training finished successfully.\n", "Telemetry data couldn't be sent, but the command ran successfully.\n", "[WARNING]: \n", "Execution status: PASS\n", "2026-02-02 17:57:16,246 [TAO Toolkit] [INFO] nvidia_tao_cli.components.docker_handler.docker_handler 363: Stopping container.\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# train model\n", "!tao model classification_tf1 train -e $TAO_SPECS_DIR/resnet50/combined_config.txt \\\n", " -r $TAO_EXPERIMENT_DIR/resnet50 \\\n", " -k $KEY" ] }, { "cell_type": "markdown", "id": "5ad44218-d756-4e76-883f-0fcc76044254", "metadata": {}, "source": [ "**Note**: The training may take some time to complete. For this dataset, each epoch takes ~20 seconds using 32 as `batch_size`. If the training process is taking too long due to a high `n_epochs`, you can interrupt the kernel to stop the training prematurely and start over with a lower `n_epochs`. The last training step (epoch) of the model is saved in the `/workspace/tao-experiments/classification/resnet50/weights` directory, which is used for evaluation, inference, and export. " ] }, { "cell_type": "markdown", "id": "dc293375-1728-49ef-a331-a7d949ee6229", "metadata": {}, "source": [ "\n", "### Model Evaluation ###\n", "The model should be evaluated for its performance at the end of training. For classification, this is typically characterized as the accuracy performance. Sometimes we need more insight into accuracy using metrics such as **precision**, **recall**, and **f1-score** for every class. To achieve this, the TAO Toolkit first produces a [**confusion matrix**](https://en.wikipedia.org/wiki/Confusion_matrix). " ] }, { "cell_type": "markdown", "id": "5ec8e101-627d-4f7b-a07f-166498d6c9f7", "metadata": {}, "source": [ "\n", "#### Exercise #4 - Modify Eval Config ####\n", "The `eval_config` describes the configurable parameters for evaluating a classification model. \n", "* `eval_dataset_path (str)`: UNIX format path to the root directory of the evaluation dataset. This can be the train, validation, or test dataset, depending on the usage. Normally, a separate test set would be used for this purpose to be independent from the training process. However, given the limited data size, we will evaluate our model based on the _validation_ dataset. \n", "* `model_path (str)`: UNIX format path to the root directory of the model file you would like to evaluate.\n", " * We can choose where to store the trained model - in this case we used `/workspace/tao-experiments/classification/resnet50` inside of the TAO container, which is mapped to `tao_project/classification/resnet50` in our local drive. Furthermore, the trained model name will follow the format `_0.hdf5`, unless specified otherwise. Therefore we should use `/workspace/tao-experiments/classification/resnet50/weights/resnet_010.hdf5` if `n_epochs` is set to `10`. \n", "* `top_k (int)`: The number of elements to look at when calculating the top-K classification categorical accuracy metric.\n", "* `batch_size (int)`: Number of images per batch when evaluating the model.\n", "* `n_workers (int)`: Number of workers fetching batches of images in the evaluation dataloader.\n", "* `enable_center_crop (bool)`: Enable center crop for input images or not. Usually this parameter is set to `True` to achieve better accuracy.\n", "\n", "**Instructions**:
\n", "* Execute the below cell to view the model trained and familiarize with how it should be referenced. \n", "* Modify the `eval_config`[(here)](tao_project/spec_files/resnet50/eval_config.txt) section of the training configuration file by changing the `` into acceptable values and **save changes**. " ] }, { "cell_type": "code", "execution_count": 28, "id": "0f92b108-f0f9-419b-b829-c05f7542c5e9", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Model for every epoch:\n", "---------------------\n", "total 1.8G\n", "-rw-r--r-- 1 root root 181M Feb 2 17:53 resnet_001.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 17:54 resnet_002.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 17:54 resnet_003.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 17:54 resnet_004.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 17:55 resnet_005.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 17:55 resnet_006.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 17:55 resnet_007.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 17:56 resnet_008.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 17:56 resnet_009.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 17:56 resnet_010.hdf5\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "print('Model for every epoch:')\n", "print('---------------------')\n", "!ls -ltrh $LOCAL_EXPERIMENT_DIR/resnet50/weights" ] }, { "cell_type": "code", "execution_count": 29, "id": "65461fd6-72d1-4365-9fca-c7d6cef0fce3", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "eval_config {\n", " eval_dataset_path: \"/workspace/tao-experiments/data/val\"\n", " model_path: \"/workspace/tao-experiments/classification/resnet50/weights/resnet_010.hdf5\"\n", " top_k: 1\n", " batch_size: 32\n", " n_workers: 8\n", " enable_center_crop: False\n", "}" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# read the config file\n", "!cat $LOCAL_SPECS_DIR/resnet50/eval_config.txt" ] }, { "cell_type": "markdown", "id": "5100f372-8ce1-41e8-99f4-72bd8d30d611", "metadata": {}, "source": [ "\n", "### Combine Configuration Files and Evaluate Model ###" ] }, { "cell_type": "code", "execution_count": 30, "id": "5975aa60-47aa-4aff-b344-9d706a972ca1", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "model_config {\n", " arch: \"resnet\"\n", " n_layers: 50\n", " freeze_blocks: 0\n", " use_batch_norm: True\n", " input_image_size: \"3,224,224\"\n", "}train_config {\n", " train_dataset_path: \"/workspace/tao-experiments/data/train\"\n", " val_dataset_path: \"/workspace/tao-experiments/data/val\"\n", " pretrained_model_path: \"/workspace/tao-experiments/classification/pretrained_resnet50/pretrained_classification_vresnet50/resnet_50.hdf5\"\n", " optimizer {\n", " sgd {\n", " lr: 0.01\n", " decay: 0.0\n", " momentum: 0.9\n", " nesterov: False\n", " }\n", " }\n", " n_epochs: 10\n", " batch_size_per_gpu: 32\n", " n_workers: 8\n", " enable_random_crop: False\n", " enable_center_crop: False\n", " enable_color_augmentation: False\n", " preprocess_mode: \"caffe\"\n", " reg_config {\n", " type: \"L2\"\n", " scope: \"Conv2D, Dense\"\n", " weight_decay: 0.00005\n", " }\n", " lr_config {\n", " step {\n", " learning_rate: 0.006\n", " step_size: 10\n", " gamma: 0.1\n", " }\n", " }\n", "}\n", "eval_config {\n", " eval_dataset_path: \"/workspace/tao-experiments/data/val\"\n", " model_path: \"/workspace/tao-experiments/classification/resnet50/weights/resnet_010.hdf5\"\n", " top_k: 1\n", " batch_size: 32\n", " n_workers: 8\n", " enable_center_crop: False\n", "}" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# combining configuration components in separate files and writing into one\n", "!cat $LOCAL_SPECS_DIR/resnet50/model_config.txt \\\n", " $LOCAL_SPECS_DIR/resnet50/train_config.txt \\\n", " $LOCAL_SPECS_DIR/resnet50/eval_config.txt \\\n", " > $LOCAL_SPECS_DIR/resnet50/combined_config.txt\n", "!cat $LOCAL_SPECS_DIR/resnet50/combined_config.txt" ] }, { "cell_type": "markdown", "id": "3aa2509c-d095-4174-bd48-369068dae547", "metadata": {}, "source": [ "We can evaluate the model with the `evaluate` subtask. " ] }, { "cell_type": "raw", "id": "fefc75a1-ce72-4f4f-9f78-ef8035bbbe03", "metadata": {}, "source": [ "tao model classification_tf1 evaluate [-h] -e \n", " -k \n", " [--gpus GPUS]\n", " [--gpu_index GPU_INDEX]" ] }, { "cell_type": "markdown", "id": "ddc9d887-c398-4a6d-8c5a-e2e5629782a7", "metadata": {}, "source": [ "The `evaluate` subtask runs evaluation on the same validation set that was used during training. We can also run evaluation on an earlier model by editing the spec file to point to the intended model. When using the `evaluate` subtask, the `-e` argument indicates the path to the spec file and the `-k` argument indicates the key to _load_ the model. " ] }, { "cell_type": "code", "execution_count": 31, "id": "83adf120-92b2-48f6-8853-78b6c3bc2068", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "2026-02-02 17:57:17,379 [TAO Toolkit] [INFO] root 160: Registry: ['nvcr.io']\n", "2026-02-02 17:57:17,467 [TAO Toolkit] [INFO] nvidia_tao_cli.components.instance_handler.local_instance 360: Running command in container: nvcr.io/nvidia/tao/tao-toolkit:5.0.0-tf1.15.5\n", "2026-02-02 17:57:17,496 [TAO Toolkit] [INFO] nvidia_tao_cli.components.docker_handler.docker_handler 301: Printing tty value True\n", "Using TensorFlow backend.\n", "2026-02-02 17:57:18.606192: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 17:57:18,663 [TAO Toolkit] [WARNING] tensorflow 40: Deprecation warnings have been disabled. Set TF_ENABLE_DEPRECATION_WARNINGS=1 to re-enable them.\n", "2026-02-02 17:57:20,810 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use sklearn by default. This improves performance in some cases. To enable sklearn export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 17:57:20,924 [TAO Toolkit] [WARNING] tensorflow 42: TensorFlow will not use Dask by default. This improves performance in some cases. To enable Dask export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 17:57:20,938 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use Pandas by default. This improves performance in some cases. To enable Pandas export the environment variable TF_ALLOW_IOLIBS=1.\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:150: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_hue(img, max_delta=10.0):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:173: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_saturation(img, max_shift):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:183: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_contrast(img, center, max_contrast_scale):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:192: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_shift(x_img, shift_stddev):\n", "2026-02-02 17:57:23.387569: I tensorflow/core/platform/profile_utils/cpu_utils.cc:109] CPU Frequency: 2499995000 Hz\n", "2026-02-02 17:57:23.387915: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x93c0c50 initialized for platform Host (this does not guarantee that XLA will be used). Devices:\n", "2026-02-02 17:57:23.387952: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version\n", "2026-02-02 17:57:23.389402: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcuda.so.1\n", "2026-02-02 17:57:23.577566: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:57:23.579528: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x91dc000 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:\n", "2026-02-02 17:57:23.579561: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Tesla T4, Compute Capability 7.5\n", "2026-02-02 17:57:23.579853: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:57:23.581634: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1674] Found device 0 with properties: \n", "name: Tesla T4 major: 7 minor: 5 memoryClockRate(GHz): 1.59\n", "pciBusID: 0000:00:1e.0\n", "2026-02-02 17:57:23.581694: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 17:57:23.581837: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcublas.so.12\n", "2026-02-02 17:57:23.583975: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcufft.so.11\n", "2026-02-02 17:57:23.584077: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcurand.so.10\n", "2026-02-02 17:57:23.588389: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcusolver.so.11\n", "2026-02-02 17:57:23.589527: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcusparse.so.12\n", "2026-02-02 17:57:23.589610: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudnn.so.8\n", "2026-02-02 17:57:23.589739: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:57:23.591576: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:57:23.593292: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1802] Adding visible gpu devices: 0\n", "2026-02-02 17:57:23.593338: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 17:57:23.602346: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1214] Device interconnect StreamExecutor with strength 1 edge matrix:\n", "2026-02-02 17:57:23.602376: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1220] 0 \n", "2026-02-02 17:57:23.602389: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1233] 0: N \n", "2026-02-02 17:57:23.602579: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:57:23.604390: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 17:57:23.606128: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1359] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 13496 MB memory) -> physical GPU (device: 0, name: Tesla T4, pci bus id: 0000:00:1e.0, compute capability: 7.5)\n", "Using TensorFlow backend.\n", "WARNING:tensorflow:Deprecation warnings have been disabled. Set TF_ENABLE_DEPRECATION_WARNINGS=1 to re-enable them.\n", "WARNING:tensorflow:TensorFlow will not use sklearn by default. This improves performance in some cases. To enable sklearn export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING: TensorFlow will not use sklearn by default. This improves performance in some cases. To enable sklearn export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING:tensorflow:TensorFlow will not use Dask by default. This improves performance in some cases. To enable Dask export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING: TensorFlow will not use Dask by default. This improves performance in some cases. To enable Dask export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING:tensorflow:TensorFlow will not use Pandas by default. This improves performance in some cases. To enable Pandas export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING: TensorFlow will not use Pandas by default. This improves performance in some cases. To enable Pandas export the environment variable TF_ALLOW_IOLIBS=1.\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:150: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_hue(img, max_delta=10.0):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:173: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_saturation(img, max_shift):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:183: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_contrast(img, center, max_contrast_scale):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:192: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_shift(x_img, shift_stddev):\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/evaluate.py:44: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/evaluate.py:44: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/evaluate.py:46: The name tf.Session is deprecated. Please use tf.compat.v1.Session instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/evaluate.py:46: The name tf.Session is deprecated. Please use tf.compat.v1.Session instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/evaluate.py:157: The name tf.logging.set_verbosity is deprecated. Please use tf.compat.v1.logging.set_verbosity instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/evaluate.py:157: The name tf.logging.set_verbosity is deprecated. Please use tf.compat.v1.logging.set_verbosity instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/evaluate.py:157: The name tf.logging.INFO is deprecated. Please use tf.compat.v1.logging.INFO instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/evaluate.py:157: The name tf.logging.INFO is deprecated. Please use tf.compat.v1.logging.INFO instead.\n", "\n", "INFO: Loading experiment spec at /workspace/tao-experiments/spec_files/resnet50/combined_config.txt.\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:245: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:245: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:1834: The name tf.nn.fused_batch_norm is deprecated. Please use tf.compat.v1.nn.fused_batch_norm instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:1834: The name tf.nn.fused_batch_norm is deprecated. Please use tf.compat.v1.nn.fused_batch_norm instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/third_party/keras/tensorflow_backend.py:199: The name tf.nn.avg_pool is deprecated. Please use tf.nn.avg_pool2d instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/third_party/keras/tensorflow_backend.py:199: The name tf.nn.avg_pool is deprecated. Please use tf.nn.avg_pool2d instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:174: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:174: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:190: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:190: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:199: The name tf.is_variable_initialized is deprecated. Please use tf.compat.v1.is_variable_initialized instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:199: The name tf.is_variable_initialized is deprecated. Please use tf.compat.v1.is_variable_initialized instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:206: The name tf.variables_initializer is deprecated. Please use tf.compat.v1.variables_initializer instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:206: The name tf.variables_initializer is deprecated. Please use tf.compat.v1.variables_initializer instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n", "\n", "__________________________________________________________________________________________________\n", "Layer (type) Output Shape Param # Connected to \n", "==================================================================================================\n", "input_1 (InputLayer) (None, 3, 224, 224) 0 \n", "__________________________________________________________________________________________________\n", "conv1 (Conv2D) (None, 64, 112, 112) 9408 input_1[0][0] \n", "__________________________________________________________________________________________________\n", "bn_conv1 (BatchNormalization) (None, 64, 112, 112) 256 conv1[0][0] \n", "__________________________________________________________________________________________________\n", "activation_1 (Activation) (None, 64, 112, 112) 0 bn_conv1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_1 (Conv2D) (None, 64, 56, 56) 4096 activation_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu_1 (Activation) (None, 64, 56, 56) 0 block_1a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu_2 (Activation) (None, 64, 56, 56) 0 block_1a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_shortcut (Conv2D) (None, 256, 56, 56) 16384 activation_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_shortcut (BatchNorm (None, 256, 56, 56) 1024 block_1a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_1 (Add) (None, 256, 56, 56) 0 block_1a_bn_3[0][0] \n", " block_1a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu (Activation) (None, 256, 56, 56) 0 add_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_1 (Conv2D) (None, 64, 56, 56) 16384 block_1a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu_1 (Activation) (None, 64, 56, 56) 0 block_1b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu_2 (Activation) (None, 64, 56, 56) 0 block_1b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_2 (Add) (None, 256, 56, 56) 0 block_1b_bn_3[0][0] \n", " block_1a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu (Activation) (None, 256, 56, 56) 0 add_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_1 (Conv2D) (None, 64, 56, 56) 16384 block_1b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu_1 (Activation) (None, 64, 56, 56) 0 block_1c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu_2 (Activation) (None, 64, 56, 56) 0 block_1c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_3 (Add) (None, 256, 56, 56) 0 block_1c_bn_3[0][0] \n", " block_1b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu (Activation) (None, 256, 56, 56) 0 add_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_1 (Conv2D) (None, 128, 28, 28) 32768 block_1c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu_1 (Activation) (None, 128, 28, 28) 0 block_2a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu_2 (Activation) (None, 128, 28, 28) 0 block_2a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_shortcut (Conv2D) (None, 512, 28, 28) 131072 block_1c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_shortcut (BatchNorm (None, 512, 28, 28) 2048 block_2a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_4 (Add) (None, 512, 28, 28) 0 block_2a_bn_3[0][0] \n", " block_2a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu (Activation) (None, 512, 28, 28) 0 add_4[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu_1 (Activation) (None, 128, 28, 28) 0 block_2b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu_2 (Activation) (None, 128, 28, 28) 0 block_2b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_5 (Add) (None, 512, 28, 28) 0 block_2b_bn_3[0][0] \n", " block_2a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu (Activation) (None, 512, 28, 28) 0 add_5[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu_1 (Activation) (None, 128, 28, 28) 0 block_2c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu_2 (Activation) (None, 128, 28, 28) 0 block_2c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_6 (Add) (None, 512, 28, 28) 0 block_2c_bn_3[0][0] \n", " block_2b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu (Activation) (None, 512, 28, 28) 0 add_6[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2d_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu_1 (Activation) (None, 128, 28, 28) 0 block_2d_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2d_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2d_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu_2 (Activation) (None, 128, 28, 28) 0 block_2d_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2d_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2d_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_7 (Add) (None, 512, 28, 28) 0 block_2d_bn_3[0][0] \n", " block_2c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu (Activation) (None, 512, 28, 28) 0 add_7[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_1 (Conv2D) (None, 256, 14, 14) 131072 block_2d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu_1 (Activation) (None, 256, 14, 14) 0 block_3a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu_2 (Activation) (None, 256, 14, 14) 0 block_3a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_shortcut (Conv2D) (None, 1024, 14, 14) 524288 block_2d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_shortcut (BatchNorm (None, 1024, 14, 14) 4096 block_3a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_8 (Add) (None, 1024, 14, 14) 0 block_3a_bn_3[0][0] \n", " block_3a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu (Activation) (None, 1024, 14, 14) 0 add_8[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu_1 (Activation) (None, 256, 14, 14) 0 block_3b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu_2 (Activation) (None, 256, 14, 14) 0 block_3b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_9 (Add) (None, 1024, 14, 14) 0 block_3b_bn_3[0][0] \n", " block_3a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu (Activation) (None, 1024, 14, 14) 0 add_9[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu_1 (Activation) (None, 256, 14, 14) 0 block_3c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu_2 (Activation) (None, 256, 14, 14) 0 block_3c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_10 (Add) (None, 1024, 14, 14) 0 block_3c_bn_3[0][0] \n", " block_3b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu (Activation) (None, 1024, 14, 14) 0 add_10[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3d_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu_1 (Activation) (None, 256, 14, 14) 0 block_3d_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3d_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3d_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu_2 (Activation) (None, 256, 14, 14) 0 block_3d_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3d_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3d_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_11 (Add) (None, 1024, 14, 14) 0 block_3d_bn_3[0][0] \n", " block_3c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu (Activation) (None, 1024, 14, 14) 0 add_11[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3e_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu_1 (Activation) (None, 256, 14, 14) 0 block_3e_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3e_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3e_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu_2 (Activation) (None, 256, 14, 14) 0 block_3e_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3e_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3e_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_12 (Add) (None, 1024, 14, 14) 0 block_3e_bn_3[0][0] \n", " block_3d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu (Activation) (None, 1024, 14, 14) 0 add_12[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3e_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3f_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu_1 (Activation) (None, 256, 14, 14) 0 block_3f_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3f_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3f_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu_2 (Activation) (None, 256, 14, 14) 0 block_3f_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3f_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3f_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_13 (Add) (None, 1024, 14, 14) 0 block_3f_bn_3[0][0] \n", " block_3e_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu (Activation) (None, 1024, 14, 14) 0 add_13[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_1 (Conv2D) (None, 512, 14, 14) 524288 block_3f_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu_1 (Activation) (None, 512, 14, 14) 0 block_4a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu_2 (Activation) (None, 512, 14, 14) 0 block_4a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_shortcut (Conv2D) (None, 2048, 14, 14) 2097152 block_3f_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_shortcut (BatchNorm (None, 2048, 14, 14) 8192 block_4a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_14 (Add) (None, 2048, 14, 14) 0 block_4a_bn_3[0][0] \n", " block_4a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu (Activation) (None, 2048, 14, 14) 0 add_14[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_1 (Conv2D) (None, 512, 14, 14) 1048576 block_4a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu_1 (Activation) (None, 512, 14, 14) 0 block_4b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu_2 (Activation) (None, 512, 14, 14) 0 block_4b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_15 (Add) (None, 2048, 14, 14) 0 block_4b_bn_3[0][0] \n", " block_4a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu (Activation) (None, 2048, 14, 14) 0 add_15[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_1 (Conv2D) (None, 512, 14, 14) 1048576 block_4b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu_1 (Activation) (None, 512, 14, 14) 0 block_4c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu_2 (Activation) (None, 512, 14, 14) 0 block_4c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_16 (Add) (None, 2048, 14, 14) 0 block_4c_bn_3[0][0] \n", " block_4b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu (Activation) (None, 2048, 14, 14) 0 add_16[0][0] \n", "__________________________________________________________________________________________________\n", "avg_pool (AveragePooling2D) (None, 2048, 1, 1) 0 block_4c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "flatten (Flatten) (None, 2048) 0 avg_pool[0][0] \n", "__________________________________________________________________________________________________\n", "predictions (Dense) (None, 2) 4098 flatten[0][0] \n", "==================================================================================================\n", "Total params: 23,565,250\n", "Trainable params: 23,502,722\n", "Non-trainable params: 62,528\n", "__________________________________________________________________________________________________\n", "Found 571 images belonging to 2 classes.\n", "INFO: Processing dataset (evaluation): /workspace/tao-experiments/data/val\n", "Evaluation Loss: 0.3445939733247206\n", "Evaluation Top K accuracy: 0.9474605954465849\n", "Found 571 images belonging to 2 classes.\n", "INFO: Calculating per-class P/R and confusion matrix. It may take a while...\n", "Confusion Matrix\n", "[[ 0 30]\n", " [ 0 541]]\n", "Classification Report\n", "/usr/local/lib/python3.8/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n", " _warn_prf(average, modifier, msg_start, len(result))\n", "/usr/local/lib/python3.8/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n", " _warn_prf(average, modifier, msg_start, len(result))\n", "/usr/local/lib/python3.8/dist-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n", " _warn_prf(average, modifier, msg_start, len(result))\n", " precision recall f1-score support\n", "\n", " defect 0.00 0.00 0.00 30\n", " notdefect 0.95 1.00 0.97 541\n", "\n", " accuracy 0.95 571\n", " macro avg 0.47 0.50 0.49 571\n", "weighted avg 0.90 0.95 0.92 571\n", "\n", "Telemetry data couldn't be sent, but the command ran successfully.\n", "[WARNING]: \n", "Execution status: PASS\n", "2026-02-02 17:58:12,311 [TAO Toolkit] [INFO] nvidia_tao_cli.components.docker_handler.docker_handler 363: Stopping container.\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# evaluate the model using the same validation set as training\n", "!tao model classification_tf1 evaluate -e $TAO_SPECS_DIR/resnet50/combined_config.txt\\\n", " -k $KEY" ] }, { "cell_type": "markdown", "id": "6c5d4db0-7b8b-4e12-9f76-19ec2c42c4ed", "metadata": {}, "source": [ "

\n", "\n", "The model that we've trained may not do very well with predicting `defect` yet. For this dataset, we have tested that allowing the model to train for at least **25 epochs** is a good starting point. In the following section, we're going to fine-tune the model to improve it's performance. " ] }, { "cell_type": "markdown", "id": "b46c8022-246c-4039-ab95-c9e57b1b7ea9", "metadata": {}, "source": [ "\n", "## Model Tuning ##" ] }, { "cell_type": "markdown", "id": "b177df96-712f-4370-bb9e-fccf4dfac59a", "metadata": {}, "source": [ "\n", "### Data Augmentation ###\n", "Deep learning models require training with vast amounts of data to achieve accurate results. DALI can not only read images from disk and batch them into tensors, it can also perform various augmentations on those images to improve deep learning training results. [Data augmentation](https://en.wikipedia.org/wiki/Data_augmentation) artificially increases the size of a dataset by introducing random disturbances to the data, such as _geometric deformations_, _color transforms_, _noise addition_, and so on. These disturbances help produce models that are more robust in their predictions, avoid overfitting, and deliver better accuracy. We will use DALI to demonstrate data augmentation that we will introduce for model training, such as _cropping_, _resizing_, and _flipping_. " ] }, { "cell_type": "code", "execution_count": 32, "id": "6d39abb5-80a8-4a3b-857d-22fff953a269", "metadata": {}, "outputs": [], "source": [ "# DO NOT CHANGE THIS CELL\n", "# import dependencies\n", "from nvidia.dali.pipeline import Pipeline\n", "from nvidia.dali import pipeline_def\n", "import nvidia.dali.fn as fn\n", "from PIL import Image\n", "import warnings\n", "\n", "warnings.filterwarnings(\"ignore\")" ] }, { "cell_type": "code", "execution_count": 33, "id": "eeb32fcb-12df-4fbf-92c7-2dbb0c8fe712", "metadata": {}, "outputs": [], "source": [ "# DO NOT CHANGE THIS CELL\n", "batch_size=8\n", "defect_label_map={'notdefect': 1, 'defect': 0}\n", "defect_inverse_map={v: k for k, v in defect_label_map.items()}\n", "defect_train_df=capacitor_df[(capacitor_df['true_defect']=='defect') & (capacitor_df['dataset']=='train')]\n", "n_iter=10\n", "\n", "@pipeline_def\n", "def augmentation_pipeline():\n", " # use fn.readers.file to read encoded images and labels from the hard drive\n", " jpgs, labels=fn.readers.file(files=defect_train_df['defect_img_path'].to_list(), labels=defect_train_df['true_defect'].map(defect_label_map).to_list())\n", " # use the fn.decoders.image operation to decode images from JPG to RGB\n", " images=fn.decoders.image(jpgs, device='cpu')\n", " # use the fn.rotate operation to rotate image\n", " rotated_images = fn.rotate(images.gpu(), angle=90, fill_value=0)\n", " return rotated_images, labels" ] }, { "cell_type": "code", "execution_count": 34, "id": "7d4f932d-c24b-4915-8415-5750b996a1e3", "metadata": {}, "outputs": [], "source": [ "# DO NOT CHANGE THIS CELL\n", "# define a function display images\n", "def save_images(image_batch, label_batch, batch_count):\n", " for idx in range(batch_size): \n", " image_ary=np.array(image_batch[idx])\n", " im=Image.fromarray(image_ary)\n", " im.save(os.path.join(os.environ['LOCAL_DATA_DIR'], 'train', defect_inverse_map[np.array(labels[idx])[0]], 'augmented_'+str(batch_count)+'_'+str(idx))+'.jpg')" ] }, { "cell_type": "code", "execution_count": 35, "id": "2716affc-5d9d-434e-9e0e-1c602e2d7ffd", "metadata": { "tags": [] }, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "/usr/local/lib/python3.10/dist-packages/nvidia/dali/backend.py:83: Warning: nvidia-dali-cuda120 is no longer shipped with CUDA runtime. You need to install it separately. cuFFT is typically provided with CUDA Toolkit installation or an appropriate wheel. Please check https://docs.nvidia.com/cuda/cuda-quick-start-guide/index.html#pip-wheels-installation-linux for the reference.\n", " deprecation_warning(\n", "/usr/local/lib/python3.10/dist-packages/nvidia/dali/backend.py:94: Warning: nvidia-dali-cuda120 is no longer shipped with CUDA runtime. You need to install it separately. NPP is typically provided with CUDA Toolkit installation or an appropriate wheel. Please check https://docs.nvidia.com/cuda/cuda-quick-start-guide/index.html#pip-wheels-installation-linux for the reference.\n", " deprecation_warning(\n", "/usr/local/lib/python3.10/dist-packages/nvidia/dali/backend.py:105: Warning: nvidia-dali-cuda120 is no longer shipped with CUDA runtime. You need to install it separately. nvJPEG is typically provided with CUDA Toolkit installation or an appropriate wheel. Please check https://docs.nvidia.com/cuda/cuda-quick-start-guide/index.html#pip-wheels-installation-linux for the reference.\n", " deprecation_warning(\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# build pipeline\n", "augmentation_pipe=augmentation_pipeline(batch_size=batch_size, num_threads=4, device_id=0)\n", "augmentation_pipe.build()" ] }, { "cell_type": "code", "execution_count": 36, "id": "99468976-7b46-458b-b411-b747e38f0109", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "It took 0.1 seconds to create 69 images.\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# time the process\n", "start=time.time()\n", "\n", "# iterate and create some augmented data\n", "for iter in range(n_iter): \n", " augmentation_pipe_output=augmentation_pipe.run()\n", " augmented_images, labels=augmentation_pipe_output\n", " save_images(augmented_images.as_cpu(), labels, iter)\n", "\n", "print('It took {} seconds to create {} images.'.format(round(time.time()-start, 2), len(defect_train_df)))" ] }, { "cell_type": "markdown", "id": "14b4291e-b2ec-43f3-a767-8949f64c1ad9", "metadata": {}, "source": [ "

\n", "\n", "DALI does not support moving the data from the GPU to the CPU within the pipeline. That is why a CPU operation cannot follow a GPU one. Since `augmentation_pipe_output` contains 2 TensorLists, but this time the first output, result of the `rotate` operation, is on the GPU. We cannot access contents of `TensorListGPU` directly from the CPU, so in order to show the results we need to copy it to the CPU by using `as_cpu` method." ] }, { "cell_type": "markdown", "id": "2b0376da-56bf-4396-a156-2c8b7ecf8bfb", "metadata": {}, "source": [ "\n", "### Exercise #5 - Retrain Model ###\n", "We will use the same configuration file to train a new model with the augmented dataset. \n", "\n", "**instructions**:
\n", "* Execute the below cell to train another model using the augmented dataset. Note the output directory is different than the model previously trained as `resnet50_aug`. \n", "* Execute the cell below to view the model trained and familiarize with how it should be referenced. \n", "* Modify the `eval_config`[(here)](tao_project/spec_files/resnet50/eval_config.txt) section of the training configuration file by changing the `model_path` to reference the newly trained model and **save changes**. \n", "* Execute the cell below to create a new configuration file. \n", "* Execute the cell below to evaluate the new model using the same validation set as before. " ] }, { "cell_type": "code", "execution_count": 37, "id": "3c67cfe2-4ccf-483c-9e42-6c0c32392168", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "2026-02-02 18:00:42,298 [TAO Toolkit] [INFO] root 160: Registry: ['nvcr.io']\n", "2026-02-02 18:00:42,389 [TAO Toolkit] [INFO] nvidia_tao_cli.components.instance_handler.local_instance 360: Running command in container: nvcr.io/nvidia/tao/tao-toolkit:5.0.0-tf1.15.5\n", "2026-02-02 18:00:42,399 [TAO Toolkit] [INFO] nvidia_tao_cli.components.docker_handler.docker_handler 301: Printing tty value True\n", "Using TensorFlow backend.\n", "2026-02-02 18:00:43.373965: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 18:00:43,427 [TAO Toolkit] [WARNING] tensorflow 40: Deprecation warnings have been disabled. Set TF_ENABLE_DEPRECATION_WARNINGS=1 to re-enable them.\n", "2026-02-02 18:00:44,579 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use sklearn by default. This improves performance in some cases. To enable sklearn export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 18:00:44,618 [TAO Toolkit] [WARNING] tensorflow 42: TensorFlow will not use Dask by default. This improves performance in some cases. To enable Dask export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 18:00:44,622 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use Pandas by default. This improves performance in some cases. To enable Pandas export the environment variable TF_ALLOW_IOLIBS=1.\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:150: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_hue(img, max_delta=10.0):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:173: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_saturation(img, max_shift):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:183: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_contrast(img, center, max_contrast_scale):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:192: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_shift(x_img, shift_stddev):\n", "2026-02-02 18:00:46.511614: I tensorflow/core/platform/profile_utils/cpu_utils.cc:109] CPU Frequency: 2499995000 Hz\n", "2026-02-02 18:00:46.511997: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x7a12010 initialized for platform Host (this does not guarantee that XLA will be used). Devices:\n", "2026-02-02 18:00:46.512035: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version\n", "2026-02-02 18:00:46.513408: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcuda.so.1\n", "2026-02-02 18:00:46.743786: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:00:46.745582: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x782d3c0 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:\n", "2026-02-02 18:00:46.745615: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Tesla T4, Compute Capability 7.5\n", "2026-02-02 18:00:46.745933: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:00:46.747518: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1674] Found device 0 with properties: \n", "name: Tesla T4 major: 7 minor: 5 memoryClockRate(GHz): 1.59\n", "pciBusID: 0000:00:1e.0\n", "2026-02-02 18:00:46.747581: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 18:00:46.747700: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcublas.so.12\n", "2026-02-02 18:00:46.749986: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcufft.so.11\n", "2026-02-02 18:00:46.750107: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcurand.so.10\n", "2026-02-02 18:00:46.753814: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcusolver.so.11\n", "2026-02-02 18:00:46.754971: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcusparse.so.12\n", "2026-02-02 18:00:46.755060: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudnn.so.8\n", "2026-02-02 18:00:46.755189: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:00:46.756865: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:00:46.758413: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1802] Adding visible gpu devices: 0\n", "2026-02-02 18:00:46.758464: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 18:00:46.766874: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1214] Device interconnect StreamExecutor with strength 1 edge matrix:\n", "2026-02-02 18:00:46.766904: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1220] 0 \n", "2026-02-02 18:00:46.766920: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1233] 0: N \n", "2026-02-02 18:00:46.767099: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:00:46.768737: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:00:46.770317: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1359] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 13332 MB memory) -> physical GPU (device: 0, name: Tesla T4, pci bus id: 0000:00:1e.0, compute capability: 7.5)\n", "Using TensorFlow backend.\n", "WARNING:tensorflow:Deprecation warnings have been disabled. Set TF_ENABLE_DEPRECATION_WARNINGS=1 to re-enable them.\n", "WARNING:tensorflow:TensorFlow will not use sklearn by default. This improves performance in some cases. To enable sklearn export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 18:00:48,282 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use sklearn by default. This improves performance in some cases. To enable sklearn export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING:tensorflow:TensorFlow will not use Dask by default. This improves performance in some cases. To enable Dask export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 18:00:48,324 [TAO Toolkit] [WARNING] tensorflow 42: TensorFlow will not use Dask by default. This improves performance in some cases. To enable Dask export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING:tensorflow:TensorFlow will not use Pandas by default. This improves performance in some cases. To enable Pandas export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 18:00:48,328 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use Pandas by default. This improves performance in some cases. To enable Pandas export the environment variable TF_ALLOW_IOLIBS=1.\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:150: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_hue(img, max_delta=10.0):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:173: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_saturation(img, max_shift):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:183: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_contrast(img, center, max_contrast_scale):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:192: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_shift(x_img, shift_stddev):\n", "2026-02-02 18:00:49,508 [TAO Toolkit] [INFO] __main__ 388: Loading experiment spec at /workspace/tao-experiments/spec_files/resnet50/combined_config.txt.\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/train.py:398: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n", "\n", "2026-02-02 18:00:49,511 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/train.py:398: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/train.py:407: The name tf.Session is deprecated. Please use tf.compat.v1.Session instead.\n", "\n", "2026-02-02 18:00:49,511 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/train.py:407: The name tf.Session is deprecated. Please use tf.compat.v1.Session instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/train.py:431: The name tf.logging.set_verbosity is deprecated. Please use tf.compat.v1.logging.set_verbosity instead.\n", "\n", "2026-02-02 18:00:49,857 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/train.py:431: The name tf.logging.set_verbosity is deprecated. Please use tf.compat.v1.logging.set_verbosity instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/train.py:431: The name tf.logging.INFO is deprecated. Please use tf.compat.v1.logging.INFO instead.\n", "\n", "2026-02-02 18:00:49,857 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/train.py:431: The name tf.logging.INFO is deprecated. Please use tf.compat.v1.logging.INFO instead.\n", "\n", "2026-02-02 18:00:49,857 [TAO Toolkit] [INFO] __main__ 478: Default image mean [103.939, 116.779, 123.68] will be used.\n", "Found 1412 images belonging to 2 classes.\n", "2026-02-02 18:00:49,884 [TAO Toolkit] [INFO] __main__ 294: Processing dataset (train): /workspace/tao-experiments/data/train\n", "Found 571 images belonging to 2 classes.\n", "2026-02-02 18:00:49,894 [TAO Toolkit] [INFO] __main__ 311: Processing dataset (validation): /workspace/tao-experiments/data/val\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n", "\n", "2026-02-02 18:00:49,894 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n", "\n", "2026-02-02 18:00:49,895 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n", "\n", "2026-02-02 18:00:49,897 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:1834: The name tf.nn.fused_batch_norm is deprecated. Please use tf.compat.v1.nn.fused_batch_norm instead.\n", "\n", "2026-02-02 18:00:49,912 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:1834: The name tf.nn.fused_batch_norm is deprecated. Please use tf.compat.v1.nn.fused_batch_norm instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n", "\n", "2026-02-02 18:00:49,918 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/third_party/keras/tensorflow_backend.py:199: The name tf.nn.avg_pool is deprecated. Please use tf.nn.avg_pool2d instead.\n", "\n", "2026-02-02 18:00:51,469 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/third_party/keras/tensorflow_backend.py:199: The name tf.nn.avg_pool is deprecated. Please use tf.nn.avg_pool2d instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:174: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n", "\n", "2026-02-02 18:00:53,170 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:174: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:190: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n", "\n", "2026-02-02 18:00:53,170 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:190: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:199: The name tf.is_variable_initialized is deprecated. Please use tf.compat.v1.is_variable_initialized instead.\n", "\n", "2026-02-02 18:00:53,170 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:199: The name tf.is_variable_initialized is deprecated. Please use tf.compat.v1.is_variable_initialized instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:206: The name tf.variables_initializer is deprecated. Please use tf.compat.v1.variables_initializer instead.\n", "\n", "2026-02-02 18:00:54,256 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:206: The name tf.variables_initializer is deprecated. Please use tf.compat.v1.variables_initializer instead.\n", "\n", "__________________________________________________________________________________________________\n", "Layer (type) Output Shape Param # Connected to \n", "==================================================================================================\n", "input_1 (InputLayer) (None, 3, 224, 224) 0 \n", "__________________________________________________________________________________________________\n", "conv1 (Conv2D) (None, 64, 112, 112) 9408 input_1[0][0] \n", "__________________________________________________________________________________________________\n", "bn_conv1 (BatchNormalization) (None, 64, 112, 112) 256 conv1[0][0] \n", "__________________________________________________________________________________________________\n", "activation_1 (Activation) (None, 64, 112, 112) 0 bn_conv1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_1 (Conv2D) (None, 64, 56, 56) 4096 activation_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu_1 (Activation) (None, 64, 56, 56) 0 block_1a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu_2 (Activation) (None, 64, 56, 56) 0 block_1a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_shortcut (Conv2D) (None, 256, 56, 56) 16384 activation_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_shortcut (BatchNorm (None, 256, 56, 56) 1024 block_1a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_1 (Add) (None, 256, 56, 56) 0 block_1a_bn_3[0][0] \n", " block_1a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu (Activation) (None, 256, 56, 56) 0 add_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_1 (Conv2D) (None, 64, 56, 56) 16384 block_1a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu_1 (Activation) (None, 64, 56, 56) 0 block_1b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu_2 (Activation) (None, 64, 56, 56) 0 block_1b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_2 (Add) (None, 256, 56, 56) 0 block_1b_bn_3[0][0] \n", " block_1a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu (Activation) (None, 256, 56, 56) 0 add_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_1 (Conv2D) (None, 64, 56, 56) 16384 block_1b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu_1 (Activation) (None, 64, 56, 56) 0 block_1c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu_2 (Activation) (None, 64, 56, 56) 0 block_1c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_3 (Add) (None, 256, 56, 56) 0 block_1c_bn_3[0][0] \n", " block_1b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu (Activation) (None, 256, 56, 56) 0 add_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_1 (Conv2D) (None, 128, 28, 28) 32768 block_1c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu_1 (Activation) (None, 128, 28, 28) 0 block_2a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu_2 (Activation) (None, 128, 28, 28) 0 block_2a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_shortcut (Conv2D) (None, 512, 28, 28) 131072 block_1c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_shortcut (BatchNorm (None, 512, 28, 28) 2048 block_2a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_4 (Add) (None, 512, 28, 28) 0 block_2a_bn_3[0][0] \n", " block_2a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu (Activation) (None, 512, 28, 28) 0 add_4[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu_1 (Activation) (None, 128, 28, 28) 0 block_2b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu_2 (Activation) (None, 128, 28, 28) 0 block_2b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_5 (Add) (None, 512, 28, 28) 0 block_2b_bn_3[0][0] \n", " block_2a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu (Activation) (None, 512, 28, 28) 0 add_5[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu_1 (Activation) (None, 128, 28, 28) 0 block_2c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu_2 (Activation) (None, 128, 28, 28) 0 block_2c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_6 (Add) (None, 512, 28, 28) 0 block_2c_bn_3[0][0] \n", " block_2b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu (Activation) (None, 512, 28, 28) 0 add_6[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2d_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu_1 (Activation) (None, 128, 28, 28) 0 block_2d_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2d_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2d_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu_2 (Activation) (None, 128, 28, 28) 0 block_2d_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2d_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2d_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_7 (Add) (None, 512, 28, 28) 0 block_2d_bn_3[0][0] \n", " block_2c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu (Activation) (None, 512, 28, 28) 0 add_7[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_1 (Conv2D) (None, 256, 14, 14) 131072 block_2d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu_1 (Activation) (None, 256, 14, 14) 0 block_3a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu_2 (Activation) (None, 256, 14, 14) 0 block_3a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_shortcut (Conv2D) (None, 1024, 14, 14) 524288 block_2d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_shortcut (BatchNorm (None, 1024, 14, 14) 4096 block_3a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_8 (Add) (None, 1024, 14, 14) 0 block_3a_bn_3[0][0] \n", " block_3a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu (Activation) (None, 1024, 14, 14) 0 add_8[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu_1 (Activation) (None, 256, 14, 14) 0 block_3b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu_2 (Activation) (None, 256, 14, 14) 0 block_3b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_9 (Add) (None, 1024, 14, 14) 0 block_3b_bn_3[0][0] \n", " block_3a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu (Activation) (None, 1024, 14, 14) 0 add_9[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu_1 (Activation) (None, 256, 14, 14) 0 block_3c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu_2 (Activation) (None, 256, 14, 14) 0 block_3c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_10 (Add) (None, 1024, 14, 14) 0 block_3c_bn_3[0][0] \n", " block_3b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu (Activation) (None, 1024, 14, 14) 0 add_10[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3d_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu_1 (Activation) (None, 256, 14, 14) 0 block_3d_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3d_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3d_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu_2 (Activation) (None, 256, 14, 14) 0 block_3d_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3d_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3d_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_11 (Add) (None, 1024, 14, 14) 0 block_3d_bn_3[0][0] \n", " block_3c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu (Activation) (None, 1024, 14, 14) 0 add_11[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3e_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu_1 (Activation) (None, 256, 14, 14) 0 block_3e_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3e_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3e_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu_2 (Activation) (None, 256, 14, 14) 0 block_3e_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3e_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3e_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_12 (Add) (None, 1024, 14, 14) 0 block_3e_bn_3[0][0] \n", " block_3d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu (Activation) (None, 1024, 14, 14) 0 add_12[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3e_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3f_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu_1 (Activation) (None, 256, 14, 14) 0 block_3f_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3f_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3f_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu_2 (Activation) (None, 256, 14, 14) 0 block_3f_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3f_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3f_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_13 (Add) (None, 1024, 14, 14) 0 block_3f_bn_3[0][0] \n", " block_3e_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu (Activation) (None, 1024, 14, 14) 0 add_13[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_1 (Conv2D) (None, 512, 14, 14) 524288 block_3f_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu_1 (Activation) (None, 512, 14, 14) 0 block_4a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu_2 (Activation) (None, 512, 14, 14) 0 block_4a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_shortcut (Conv2D) (None, 2048, 14, 14) 2097152 block_3f_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_shortcut (BatchNorm (None, 2048, 14, 14) 8192 block_4a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_14 (Add) (None, 2048, 14, 14) 0 block_4a_bn_3[0][0] \n", " block_4a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu (Activation) (None, 2048, 14, 14) 0 add_14[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_1 (Conv2D) (None, 512, 14, 14) 1048576 block_4a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu_1 (Activation) (None, 512, 14, 14) 0 block_4b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu_2 (Activation) (None, 512, 14, 14) 0 block_4b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_15 (Add) (None, 2048, 14, 14) 0 block_4b_bn_3[0][0] \n", " block_4a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu (Activation) (None, 2048, 14, 14) 0 add_15[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_1 (Conv2D) (None, 512, 14, 14) 1048576 block_4b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu_1 (Activation) (None, 512, 14, 14) 0 block_4c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu_2 (Activation) (None, 512, 14, 14) 0 block_4c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_16 (Add) (None, 2048, 14, 14) 0 block_4c_bn_3[0][0] \n", " block_4b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu (Activation) (None, 2048, 14, 14) 0 add_16[0][0] \n", "__________________________________________________________________________________________________\n", "avg_pool (AveragePooling2D) (None, 2048, 1, 1) 0 block_4c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "flatten (Flatten) (None, 2048) 0 avg_pool[0][0] \n", "__________________________________________________________________________________________________\n", "predictions (Dense) (None, 2) 4098 flatten[0][0] \n", "==================================================================================================\n", "Total params: 23,565,250\n", "Trainable params: 23,502,722\n", "Non-trainable params: 62,528\n", "__________________________________________________________________________________________________\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n", "\n", "2026-02-02 18:02:03,625 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n", "\n", "2026-02-02 18:02:03,637 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/common/utils.py:1133: The name tf.summary.FileWriter is deprecated. Please use tf.compat.v1.summary.FileWriter instead.\n", "\n", "2026-02-02 18:02:03,662 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/common/utils.py:1133: The name tf.summary.FileWriter is deprecated. Please use tf.compat.v1.summary.FileWriter instead.\n", "\n", "2026-02-02 18:02:03,662 [TAO Toolkit] [INFO] nvidia_tao_tf1.cv.common.logging.logging 197: Log file already exists at /workspace/tao-experiments/classification/resnet50_aug/status.json\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:986: The name tf.assign_add is deprecated. Please use tf.compat.v1.assign_add instead.\n", "\n", "2026-02-02 18:02:06,793 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:986: The name tf.assign_add is deprecated. Please use tf.compat.v1.assign_add instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:973: The name tf.assign is deprecated. Please use tf.compat.v1.assign instead.\n", "\n", "2026-02-02 18:02:07,119 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:973: The name tf.assign is deprecated. Please use tf.compat.v1.assign instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/common/utils.py:1181: The name tf.summary.merge_all is deprecated. Please use tf.compat.v1.summary.merge_all instead.\n", "\n", "2026-02-02 18:02:09,662 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/common/utils.py:1181: The name tf.summary.merge_all is deprecated. Please use tf.compat.v1.summary.merge_all instead.\n", "\n", "2026-02-02 18:02:11,465 [TAO Toolkit] [INFO] root 2102: Starting Training Loop.\n", "Epoch 1/10\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/common/utils.py:199: The name tf.Summary is deprecated. Please use tf.compat.v1.Summary instead.\n", "\n", "2026-02-02 18:02:36,889 [TAO Toolkit] [WARNING] tensorflow 137: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/common/utils.py:199: The name tf.Summary is deprecated. Please use tf.compat.v1.Summary instead.\n", "\n", "45/45 [==============================] - 55s 1s/step - loss: 0.4875 - acc: 0.8917 - val_loss: 0.3248 - val_acc: 0.9475\n", "2026-02-02 18:03:21,206 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 2/10\n", "45/45 [==============================] - 20s 439ms/step - loss: 0.4047 - acc: 0.8870 - val_loss: 0.3209 - val_acc: 0.9475\n", "2026-02-02 18:03:48,274 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 3/10\n", "45/45 [==============================] - 20s 440ms/step - loss: 0.3737 - acc: 0.8965 - val_loss: 0.3221 - val_acc: 0.9492\n", "2026-02-02 18:04:09,479 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 4/10\n", "45/45 [==============================] - 20s 439ms/step - loss: 0.3694 - acc: 0.8965 - val_loss: 0.3208 - val_acc: 0.9475\n", "2026-02-02 18:04:29,688 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 5/10\n", "45/45 [==============================] - 20s 440ms/step - loss: 0.3731 - acc: 0.8972 - val_loss: 0.3189 - val_acc: 0.9475\n", "2026-02-02 18:04:49,965 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 6/10\n", "45/45 [==============================] - 20s 441ms/step - loss: 0.3704 - acc: 0.8972 - val_loss: 0.3198 - val_acc: 0.9475\n", "2026-02-02 18:05:10,253 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 7/10\n", "45/45 [==============================] - 20s 440ms/step - loss: 0.3732 - acc: 0.8965 - val_loss: 0.3193 - val_acc: 0.9492\n", "2026-02-02 18:05:30,523 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 8/10\n", "45/45 [==============================] - 20s 441ms/step - loss: 0.3788 - acc: 0.8917 - val_loss: 0.3214 - val_acc: 0.9492\n", "2026-02-02 18:05:50,831 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 9/10\n", "45/45 [==============================] - 20s 441ms/step - loss: 0.3732 - acc: 0.8965 - val_loss: 0.3227 - val_acc: 0.9492\n", "2026-02-02 18:06:11,157 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "Epoch 10/10\n", "45/45 [==============================] - 20s 441ms/step - loss: 0.3706 - acc: 0.8965 - val_loss: 0.3196 - val_acc: 0.9492\n", "2026-02-02 18:06:31,447 [TAO Toolkit] [INFO] root 2102: Training loop in progress\n", "2026-02-02 18:06:31,463 [TAO Toolkit] [INFO] root 2102: Training loop complete.\n", "2026-02-02 18:06:31,479 [TAO Toolkit] [INFO] root 2102: Final model evaluation in progress.\n", "2026-02-02 18:06:35,330 [TAO Toolkit] [INFO] root 2102: Model evaluation is complete.\n", "2026-02-02 18:06:35,330 [TAO Toolkit] [INFO] __main__ 625: Total Val Loss: 0.3195776343345642\n", "2026-02-02 18:06:35,330 [TAO Toolkit] [INFO] __main__ 626: Total Val accuracy: 0.9492118954658508\n", "2026-02-02 18:06:35,330 [TAO Toolkit] [INFO] root 2102: Training finished successfully.\n", "2026-02-02 18:06:35,330 [TAO Toolkit] [INFO] __main__ 651: Training finished successfully.\n", "Telemetry data couldn't be sent, but the command ran successfully.\n", "[WARNING]: \n", "Execution status: PASS\n", "2026-02-02 18:06:57,416 [TAO Toolkit] [INFO] nvidia_tao_cli.components.docker_handler.docker_handler 363: Stopping container.\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# train model\n", "!tao model classification_tf1 train -e $TAO_SPECS_DIR/resnet50/combined_config.txt \\\n", " -r $TAO_EXPERIMENT_DIR/resnet50_aug \\\n", " -k $KEY" ] }, { "cell_type": "code", "execution_count": 38, "id": "44bccbb3-727b-4f37-9754-0256f7c8b477", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Model for every epoch:\n", "---------------------\n", "total 1.8G\n", "-rw-r--r-- 1 root root 181M Feb 2 18:03 resnet_001.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:03 resnet_002.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:04 resnet_003.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:04 resnet_004.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:04 resnet_005.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:05 resnet_006.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:05 resnet_007.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:05 resnet_008.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:06 resnet_009.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:06 resnet_010.hdf5\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "print('Model for every epoch:')\n", "print('---------------------')\n", "!ls -ltrh $LOCAL_EXPERIMENT_DIR/resnet50_aug/weights" ] }, { "cell_type": "code", "execution_count": 39, "id": "9bac1372-7232-4fdf-86a7-cfc0c3ad2425", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "eval_config {\n", " eval_dataset_path: \"/workspace/tao-experiments/data/val\"\n", " model_path: \"/workspace/tao-experiments/classification/resnet50_aug/weights/resnet_010.hdf5\"\n", " top_k: 1\n", " batch_size: 32\n", " n_workers: 8\n", " enable_center_crop: False\n", "}" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# read the config file\n", "!cat $LOCAL_SPECS_DIR/resnet50/eval_config.txt" ] }, { "cell_type": "code", "execution_count": 40, "id": "6e01f679-513a-4761-b287-902960deaf8e", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "model_config {\n", " arch: \"resnet\"\n", " n_layers: 50\n", " freeze_blocks: 0\n", " use_batch_norm: True\n", " input_image_size: \"3,224,224\"\n", "}train_config {\n", " train_dataset_path: \"/workspace/tao-experiments/data/train\"\n", " val_dataset_path: \"/workspace/tao-experiments/data/val\"\n", " pretrained_model_path: \"/workspace/tao-experiments/classification/pretrained_resnet50/pretrained_classification_vresnet50/resnet_50.hdf5\"\n", " optimizer {\n", " sgd {\n", " lr: 0.01\n", " decay: 0.0\n", " momentum: 0.9\n", " nesterov: False\n", " }\n", " }\n", " n_epochs: 10\n", " batch_size_per_gpu: 32\n", " n_workers: 8\n", " enable_random_crop: False\n", " enable_center_crop: False\n", " enable_color_augmentation: False\n", " preprocess_mode: \"caffe\"\n", " reg_config {\n", " type: \"L2\"\n", " scope: \"Conv2D, Dense\"\n", " weight_decay: 0.00005\n", " }\n", " lr_config {\n", " step {\n", " learning_rate: 0.006\n", " step_size: 10\n", " gamma: 0.1\n", " }\n", " }\n", "}\n", "eval_config {\n", " eval_dataset_path: \"/workspace/tao-experiments/data/val\"\n", " model_path: \"/workspace/tao-experiments/classification/resnet50_aug/weights/resnet_010.hdf5\"\n", " top_k: 1\n", " batch_size: 32\n", " n_workers: 8\n", " enable_center_crop: False\n", "}" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# combining configuration components in separate files and writing into one\n", "!cat $LOCAL_SPECS_DIR/resnet50/model_config.txt \\\n", " $LOCAL_SPECS_DIR/resnet50/train_config.txt \\\n", " $LOCAL_SPECS_DIR/resnet50/eval_config.txt \\\n", " > $LOCAL_SPECS_DIR/resnet50/combined_config_aug.txt\n", "!cat $LOCAL_SPECS_DIR/resnet50/combined_config_aug.txt" ] }, { "cell_type": "code", "execution_count": 41, "id": "78e54c5f-524a-47a0-9858-315e4916066e", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "2026-02-02 18:06:58,613 [TAO Toolkit] [INFO] root 160: Registry: ['nvcr.io']\n", "2026-02-02 18:06:58,703 [TAO Toolkit] [INFO] nvidia_tao_cli.components.instance_handler.local_instance 360: Running command in container: nvcr.io/nvidia/tao/tao-toolkit:5.0.0-tf1.15.5\n", "2026-02-02 18:06:58,714 [TAO Toolkit] [INFO] nvidia_tao_cli.components.docker_handler.docker_handler 301: Printing tty value True\n", "Using TensorFlow backend.\n", "2026-02-02 18:07:00.071444: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 18:07:00,144 [TAO Toolkit] [WARNING] tensorflow 40: Deprecation warnings have been disabled. Set TF_ENABLE_DEPRECATION_WARNINGS=1 to re-enable them.\n", "2026-02-02 18:07:02,421 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use sklearn by default. This improves performance in some cases. To enable sklearn export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 18:07:02,542 [TAO Toolkit] [WARNING] tensorflow 42: TensorFlow will not use Dask by default. This improves performance in some cases. To enable Dask export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 18:07:02,556 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use Pandas by default. This improves performance in some cases. To enable Pandas export the environment variable TF_ALLOW_IOLIBS=1.\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:150: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_hue(img, max_delta=10.0):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:173: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_saturation(img, max_shift):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:183: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_contrast(img, center, max_contrast_scale):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:192: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_shift(x_img, shift_stddev):\n", "2026-02-02 18:07:05.039490: I tensorflow/core/platform/profile_utils/cpu_utils.cc:109] CPU Frequency: 2499995000 Hz\n", "2026-02-02 18:07:05.039927: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x95b8a40 initialized for platform Host (this does not guarantee that XLA will be used). Devices:\n", "2026-02-02 18:07:05.039957: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version\n", "2026-02-02 18:07:05.041294: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcuda.so.1\n", "2026-02-02 18:07:05.245009: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:07:05.247011: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x93d3df0 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:\n", "2026-02-02 18:07:05.247044: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Tesla T4, Compute Capability 7.5\n", "2026-02-02 18:07:05.247341: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:07:05.249178: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1674] Found device 0 with properties: \n", "name: Tesla T4 major: 7 minor: 5 memoryClockRate(GHz): 1.59\n", "pciBusID: 0000:00:1e.0\n", "2026-02-02 18:07:05.249238: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 18:07:05.249377: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcublas.so.12\n", "2026-02-02 18:07:05.251536: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcufft.so.11\n", "2026-02-02 18:07:05.251652: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcurand.so.10\n", "2026-02-02 18:07:05.255213: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcusolver.so.11\n", "2026-02-02 18:07:05.256388: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcusparse.so.12\n", "2026-02-02 18:07:05.256473: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudnn.so.8\n", "2026-02-02 18:07:05.256606: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:07:05.258503: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:07:05.260280: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1802] Adding visible gpu devices: 0\n", "2026-02-02 18:07:05.260330: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 18:07:05.268998: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1214] Device interconnect StreamExecutor with strength 1 edge matrix:\n", "2026-02-02 18:07:05.269027: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1220] 0 \n", "2026-02-02 18:07:05.269041: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1233] 0: N \n", "2026-02-02 18:07:05.269207: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:07:05.271079: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:07:05.272915: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1359] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 13332 MB memory) -> physical GPU (device: 0, name: Tesla T4, pci bus id: 0000:00:1e.0, compute capability: 7.5)\n", "Using TensorFlow backend.\n", "WARNING:tensorflow:Deprecation warnings have been disabled. Set TF_ENABLE_DEPRECATION_WARNINGS=1 to re-enable them.\n", "WARNING:tensorflow:TensorFlow will not use sklearn by default. This improves performance in some cases. To enable sklearn export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING: TensorFlow will not use sklearn by default. This improves performance in some cases. To enable sklearn export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING:tensorflow:TensorFlow will not use Dask by default. This improves performance in some cases. To enable Dask export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING: TensorFlow will not use Dask by default. This improves performance in some cases. To enable Dask export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING:tensorflow:TensorFlow will not use Pandas by default. This improves performance in some cases. To enable Pandas export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING: TensorFlow will not use Pandas by default. This improves performance in some cases. To enable Pandas export the environment variable TF_ALLOW_IOLIBS=1.\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:150: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_hue(img, max_delta=10.0):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:173: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_saturation(img, max_shift):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:183: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_contrast(img, center, max_contrast_scale):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:192: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_shift(x_img, shift_stddev):\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/evaluate.py:44: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/evaluate.py:44: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/evaluate.py:46: The name tf.Session is deprecated. Please use tf.compat.v1.Session instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/evaluate.py:46: The name tf.Session is deprecated. Please use tf.compat.v1.Session instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/evaluate.py:157: The name tf.logging.set_verbosity is deprecated. Please use tf.compat.v1.logging.set_verbosity instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/evaluate.py:157: The name tf.logging.set_verbosity is deprecated. Please use tf.compat.v1.logging.set_verbosity instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/evaluate.py:157: The name tf.logging.INFO is deprecated. Please use tf.compat.v1.logging.INFO instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/scripts/evaluate.py:157: The name tf.logging.INFO is deprecated. Please use tf.compat.v1.logging.INFO instead.\n", "\n", "INFO: Loading experiment spec at /workspace/tao-experiments/spec_files/resnet50/combined_config_aug.txt.\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:245: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:245: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:1834: The name tf.nn.fused_batch_norm is deprecated. Please use tf.compat.v1.nn.fused_batch_norm instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:1834: The name tf.nn.fused_batch_norm is deprecated. Please use tf.compat.v1.nn.fused_batch_norm instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/third_party/keras/tensorflow_backend.py:199: The name tf.nn.avg_pool is deprecated. Please use tf.nn.avg_pool2d instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/third_party/keras/tensorflow_backend.py:199: The name tf.nn.avg_pool is deprecated. Please use tf.nn.avg_pool2d instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:174: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:174: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:190: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:190: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:199: The name tf.is_variable_initialized is deprecated. Please use tf.compat.v1.is_variable_initialized instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:199: The name tf.is_variable_initialized is deprecated. Please use tf.compat.v1.is_variable_initialized instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:206: The name tf.variables_initializer is deprecated. Please use tf.compat.v1.variables_initializer instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:206: The name tf.variables_initializer is deprecated. Please use tf.compat.v1.variables_initializer instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n", "\n", "WARNING:tensorflow:From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n", "\n", "__________________________________________________________________________________________________\n", "Layer (type) Output Shape Param # Connected to \n", "==================================================================================================\n", "input_1 (InputLayer) (None, 3, 224, 224) 0 \n", "__________________________________________________________________________________________________\n", "conv1 (Conv2D) (None, 64, 112, 112) 9408 input_1[0][0] \n", "__________________________________________________________________________________________________\n", "bn_conv1 (BatchNormalization) (None, 64, 112, 112) 256 conv1[0][0] \n", "__________________________________________________________________________________________________\n", "activation_1 (Activation) (None, 64, 112, 112) 0 bn_conv1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_1 (Conv2D) (None, 64, 56, 56) 4096 activation_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu_1 (Activation) (None, 64, 56, 56) 0 block_1a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu_2 (Activation) (None, 64, 56, 56) 0 block_1a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_shortcut (Conv2D) (None, 256, 56, 56) 16384 activation_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_shortcut (BatchNorm (None, 256, 56, 56) 1024 block_1a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_1 (Add) (None, 256, 56, 56) 0 block_1a_bn_3[0][0] \n", " block_1a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu (Activation) (None, 256, 56, 56) 0 add_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_1 (Conv2D) (None, 64, 56, 56) 16384 block_1a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu_1 (Activation) (None, 64, 56, 56) 0 block_1b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu_2 (Activation) (None, 64, 56, 56) 0 block_1b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_2 (Add) (None, 256, 56, 56) 0 block_1b_bn_3[0][0] \n", " block_1a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu (Activation) (None, 256, 56, 56) 0 add_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_1 (Conv2D) (None, 64, 56, 56) 16384 block_1b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu_1 (Activation) (None, 64, 56, 56) 0 block_1c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu_2 (Activation) (None, 64, 56, 56) 0 block_1c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_3 (Add) (None, 256, 56, 56) 0 block_1c_bn_3[0][0] \n", " block_1b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu (Activation) (None, 256, 56, 56) 0 add_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_1 (Conv2D) (None, 128, 28, 28) 32768 block_1c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu_1 (Activation) (None, 128, 28, 28) 0 block_2a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu_2 (Activation) (None, 128, 28, 28) 0 block_2a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_shortcut (Conv2D) (None, 512, 28, 28) 131072 block_1c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_shortcut (BatchNorm (None, 512, 28, 28) 2048 block_2a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_4 (Add) (None, 512, 28, 28) 0 block_2a_bn_3[0][0] \n", " block_2a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu (Activation) (None, 512, 28, 28) 0 add_4[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu_1 (Activation) (None, 128, 28, 28) 0 block_2b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu_2 (Activation) (None, 128, 28, 28) 0 block_2b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_5 (Add) (None, 512, 28, 28) 0 block_2b_bn_3[0][0] \n", " block_2a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu (Activation) (None, 512, 28, 28) 0 add_5[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu_1 (Activation) (None, 128, 28, 28) 0 block_2c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu_2 (Activation) (None, 128, 28, 28) 0 block_2c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_6 (Add) (None, 512, 28, 28) 0 block_2c_bn_3[0][0] \n", " block_2b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu (Activation) (None, 512, 28, 28) 0 add_6[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2d_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu_1 (Activation) (None, 128, 28, 28) 0 block_2d_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2d_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2d_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu_2 (Activation) (None, 128, 28, 28) 0 block_2d_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2d_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2d_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_7 (Add) (None, 512, 28, 28) 0 block_2d_bn_3[0][0] \n", " block_2c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu (Activation) (None, 512, 28, 28) 0 add_7[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_1 (Conv2D) (None, 256, 14, 14) 131072 block_2d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu_1 (Activation) (None, 256, 14, 14) 0 block_3a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu_2 (Activation) (None, 256, 14, 14) 0 block_3a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_shortcut (Conv2D) (None, 1024, 14, 14) 524288 block_2d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_shortcut (BatchNorm (None, 1024, 14, 14) 4096 block_3a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_8 (Add) (None, 1024, 14, 14) 0 block_3a_bn_3[0][0] \n", " block_3a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu (Activation) (None, 1024, 14, 14) 0 add_8[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu_1 (Activation) (None, 256, 14, 14) 0 block_3b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu_2 (Activation) (None, 256, 14, 14) 0 block_3b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_9 (Add) (None, 1024, 14, 14) 0 block_3b_bn_3[0][0] \n", " block_3a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu (Activation) (None, 1024, 14, 14) 0 add_9[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu_1 (Activation) (None, 256, 14, 14) 0 block_3c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu_2 (Activation) (None, 256, 14, 14) 0 block_3c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_10 (Add) (None, 1024, 14, 14) 0 block_3c_bn_3[0][0] \n", " block_3b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu (Activation) (None, 1024, 14, 14) 0 add_10[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3d_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu_1 (Activation) (None, 256, 14, 14) 0 block_3d_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3d_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3d_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu_2 (Activation) (None, 256, 14, 14) 0 block_3d_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3d_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3d_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_11 (Add) (None, 1024, 14, 14) 0 block_3d_bn_3[0][0] \n", " block_3c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu (Activation) (None, 1024, 14, 14) 0 add_11[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3e_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu_1 (Activation) (None, 256, 14, 14) 0 block_3e_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3e_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3e_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu_2 (Activation) (None, 256, 14, 14) 0 block_3e_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3e_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3e_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_12 (Add) (None, 1024, 14, 14) 0 block_3e_bn_3[0][0] \n", " block_3d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu (Activation) (None, 1024, 14, 14) 0 add_12[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3e_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3f_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu_1 (Activation) (None, 256, 14, 14) 0 block_3f_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3f_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3f_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu_2 (Activation) (None, 256, 14, 14) 0 block_3f_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3f_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3f_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_13 (Add) (None, 1024, 14, 14) 0 block_3f_bn_3[0][0] \n", " block_3e_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu (Activation) (None, 1024, 14, 14) 0 add_13[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_1 (Conv2D) (None, 512, 14, 14) 524288 block_3f_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu_1 (Activation) (None, 512, 14, 14) 0 block_4a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu_2 (Activation) (None, 512, 14, 14) 0 block_4a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_shortcut (Conv2D) (None, 2048, 14, 14) 2097152 block_3f_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_shortcut (BatchNorm (None, 2048, 14, 14) 8192 block_4a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_14 (Add) (None, 2048, 14, 14) 0 block_4a_bn_3[0][0] \n", " block_4a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu (Activation) (None, 2048, 14, 14) 0 add_14[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_1 (Conv2D) (None, 512, 14, 14) 1048576 block_4a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu_1 (Activation) (None, 512, 14, 14) 0 block_4b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu_2 (Activation) (None, 512, 14, 14) 0 block_4b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_15 (Add) (None, 2048, 14, 14) 0 block_4b_bn_3[0][0] \n", " block_4a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu (Activation) (None, 2048, 14, 14) 0 add_15[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_1 (Conv2D) (None, 512, 14, 14) 1048576 block_4b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu_1 (Activation) (None, 512, 14, 14) 0 block_4c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu_2 (Activation) (None, 512, 14, 14) 0 block_4c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_16 (Add) (None, 2048, 14, 14) 0 block_4c_bn_3[0][0] \n", " block_4b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu (Activation) (None, 2048, 14, 14) 0 add_16[0][0] \n", "__________________________________________________________________________________________________\n", "avg_pool (AveragePooling2D) (None, 2048, 1, 1) 0 block_4c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "flatten (Flatten) (None, 2048) 0 avg_pool[0][0] \n", "__________________________________________________________________________________________________\n", "predictions (Dense) (None, 2) 4098 flatten[0][0] \n", "==================================================================================================\n", "Total params: 23,565,250\n", "Trainable params: 23,502,722\n", "Non-trainable params: 62,528\n", "__________________________________________________________________________________________________\n", "Found 571 images belonging to 2 classes.\n", "INFO: Processing dataset (evaluation): /workspace/tao-experiments/data/val\n", "Evaluation Loss: 0.31957761622352066\n", "Evaluation Top K accuracy: 0.9492119089316988\n", "Found 571 images belonging to 2 classes.\n", "INFO: Calculating per-class P/R and confusion matrix. It may take a while...\n", "Confusion Matrix\n", "[[ 1 29]\n", " [ 0 541]]\n", "Classification Report\n", " precision recall f1-score support\n", "\n", " defect 1.00 0.03 0.06 30\n", " notdefect 0.95 1.00 0.97 541\n", "\n", " accuracy 0.95 571\n", " macro avg 0.97 0.52 0.52 571\n", "weighted avg 0.95 0.95 0.93 571\n", "\n", "Telemetry data couldn't be sent, but the command ran successfully.\n", "[WARNING]: \n", "Execution status: PASS\n", "2026-02-02 18:07:53,833 [TAO Toolkit] [INFO] nvidia_tao_cli.components.docker_handler.docker_handler 363: Stopping container.\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# evaluate the model using the same validation set as training\n", "!tao model classification_tf1 evaluate -e $TAO_SPECS_DIR/resnet50/combined_config_aug.txt\\\n", " -k $KEY" ] }, { "cell_type": "markdown", "id": "e3d60ff5-ef94-47ef-8ecf-550ff667836c", "metadata": {}, "source": [ "\n", "### Model Inference ###\n", "The `inference` subtask for `classification_tf1` may be used to generate inference results on a directory of images. " ] }, { "cell_type": "raw", "id": "dda361eb-96b0-4cde-9057-777445e4fca4", "metadata": { "tags": [] }, "source": [ "tao model classification_tf1 inference [-h] -e \n", " -m \n", " -d \n", " -cm \n", " -k \n", " [-i IMAGE]\n", " [--gpus GPUS]\n", " [--gpu_index GPU_INDEX]" ] }, { "cell_type": "markdown", "id": "0b31b270-0704-4e96-b835-b8eec824d8f1", "metadata": {}, "source": [ "We will perform inference on non-defective images in `/workspace/tao-experiments/data/val/notdefect` as well as defective images in `/workspace/tao-experiments/data/val/defect`. When using the `inference` subtask, the `-e` argument indicates the path to the inference spec file, the `-m` argument indicates the path to the model file, the `-d` argument indicates the path to the images directory, the `-cm` argument indicates _json_ file that specifies the class index and label mapping, and the `-k` argument indicates the key to _load_ the model. " ] }, { "cell_type": "markdown", "id": "4b20cef4-6e14-45ca-86d9-28158221492b", "metadata": {}, "source": [ "\n", "### Exercise #6 - Model Inference ###\n", "\n", "**instructions**:
\n", "* Execute the below cell to view the model trained and familiarize with how it should be referenced. \n", "* Modify the `` only to reference the classification model and execute the cell below to perform model inference on non-defective images. \n", "* Modify the `` only to reference the classification model and execute the cell below to perform model inference on defective images. " ] }, { "cell_type": "code", "execution_count": 42, "id": "48f89110-f2aa-4779-8b97-8c5f772607d1", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Model for every epoch:\n", "---------------------\n", "total 1.8G\n", "-rw-r--r-- 1 root root 181M Feb 2 18:03 resnet_001.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:03 resnet_002.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:04 resnet_003.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:04 resnet_004.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:04 resnet_005.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:05 resnet_006.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:05 resnet_007.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:05 resnet_008.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:06 resnet_009.hdf5\n", "-rw-r--r-- 1 root root 181M Feb 2 18:06 resnet_010.hdf5\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "print('Model for every epoch:')\n", "print('---------------------')\n", "!ls -ltrh $LOCAL_EXPERIMENT_DIR/resnet50_aug/weights" ] }, { "cell_type": "code", "execution_count": 43, "id": "0c546165-e664-4167-ac70-f6a1a3a4d3f6", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "2026-02-02 18:07:54,597 [TAO Toolkit] [INFO] root 160: Registry: ['nvcr.io']\n", "2026-02-02 18:07:54,695 [TAO Toolkit] [INFO] nvidia_tao_cli.components.instance_handler.local_instance 360: Running command in container: nvcr.io/nvidia/tao/tao-toolkit:5.0.0-tf1.15.5\n", "2026-02-02 18:07:54,706 [TAO Toolkit] [INFO] nvidia_tao_cli.components.docker_handler.docker_handler 301: Printing tty value True\n", "Using TensorFlow backend.\n", "2026-02-02 18:07:55.541032: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 18:07:55,592 [TAO Toolkit] [WARNING] tensorflow 40: Deprecation warnings have been disabled. Set TF_ENABLE_DEPRECATION_WARNINGS=1 to re-enable them.\n", "2026-02-02 18:07:56,669 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use sklearn by default. This improves performance in some cases. To enable sklearn export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 18:07:56,723 [TAO Toolkit] [WARNING] tensorflow 42: TensorFlow will not use Dask by default. This improves performance in some cases. To enable Dask export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 18:07:56,729 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use Pandas by default. This improves performance in some cases. To enable Pandas export the environment variable TF_ALLOW_IOLIBS=1.\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:150: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_hue(img, max_delta=10.0):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:173: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_saturation(img, max_shift):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:183: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_contrast(img, center, max_contrast_scale):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:192: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_shift(x_img, shift_stddev):\n", "2026-02-02 18:07:58.621763: I tensorflow/core/platform/profile_utils/cpu_utils.cc:109] CPU Frequency: 2499995000 Hz\n", "2026-02-02 18:07:58.622126: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x95b95d0 initialized for platform Host (this does not guarantee that XLA will be used). Devices:\n", "2026-02-02 18:07:58.622161: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version\n", "2026-02-02 18:07:58.623623: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcuda.so.1\n", "2026-02-02 18:07:58.813704: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:07:58.815553: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x93d4980 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:\n", "2026-02-02 18:07:58.815586: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Tesla T4, Compute Capability 7.5\n", "2026-02-02 18:07:58.816004: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:07:58.817667: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1674] Found device 0 with properties: \n", "name: Tesla T4 major: 7 minor: 5 memoryClockRate(GHz): 1.59\n", "pciBusID: 0000:00:1e.0\n", "2026-02-02 18:07:58.817724: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 18:07:58.817834: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcublas.so.12\n", "2026-02-02 18:07:58.820056: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcufft.so.11\n", "2026-02-02 18:07:58.820163: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcurand.so.10\n", "2026-02-02 18:07:58.823896: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcusolver.so.11\n", "2026-02-02 18:07:58.825127: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcusparse.so.12\n", "2026-02-02 18:07:58.825213: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudnn.so.8\n", "2026-02-02 18:07:58.825365: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:07:58.827093: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:07:58.828761: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1802] Adding visible gpu devices: 0\n", "2026-02-02 18:07:58.828819: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 18:07:58.837512: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1214] Device interconnect StreamExecutor with strength 1 edge matrix:\n", "2026-02-02 18:07:58.837542: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1220] 0 \n", "2026-02-02 18:07:58.837558: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1233] 0: N \n", "2026-02-02 18:07:58.837750: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:07:58.839503: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:07:58.841162: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1359] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 13332 MB memory) -> physical GPU (device: 0, name: Tesla T4, pci bus id: 0000:00:1e.0, compute capability: 7.5)\n", "Using TensorFlow backend.\n", "WARNING: Deprecation warnings have been disabled. Set TF_ENABLE_DEPRECATION_WARNINGS=1 to re-enable them.\n", "WARNING: TensorFlow will not use sklearn by default. This improves performance in some cases. To enable sklearn export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING: TensorFlow will not use Dask by default. This improves performance in some cases. To enable Dask export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING: TensorFlow will not use Pandas by default. This improves performance in some cases. To enable Pandas export the environment variable TF_ALLOW_IOLIBS=1.\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:150: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_hue(img, max_delta=10.0):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:173: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_saturation(img, max_shift):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:183: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_contrast(img, center, max_contrast_scale):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:192: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_shift(x_img, shift_stddev):\n", "INFO: Loading experiment spec at /workspace/tao-experiments/spec_files/resnet50/combined_config_aug.txt.\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:245: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:1834: The name tf.nn.fused_batch_norm is deprecated. Please use tf.compat.v1.nn.fused_batch_norm instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/third_party/keras/tensorflow_backend.py:199: The name tf.nn.avg_pool is deprecated. Please use tf.nn.avg_pool2d instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:174: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:181: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:186: The name tf.Session is deprecated. Please use tf.compat.v1.Session instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:190: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:199: The name tf.is_variable_initialized is deprecated. Please use tf.compat.v1.is_variable_initialized instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:206: The name tf.variables_initializer is deprecated. Please use tf.compat.v1.variables_initializer instead.\n", "\n", "__________________________________________________________________________________________________\n", "Layer (type) Output Shape Param # Connected to \n", "==================================================================================================\n", "input_1 (InputLayer) (None, 3, 224, 224) 0 \n", "__________________________________________________________________________________________________\n", "conv1 (Conv2D) (None, 64, 112, 112) 9408 input_1[0][0] \n", "__________________________________________________________________________________________________\n", "bn_conv1 (BatchNormalization) (None, 64, 112, 112) 256 conv1[0][0] \n", "__________________________________________________________________________________________________\n", "activation_1 (Activation) (None, 64, 112, 112) 0 bn_conv1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_1 (Conv2D) (None, 64, 56, 56) 4096 activation_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu_1 (Activation) (None, 64, 56, 56) 0 block_1a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu_2 (Activation) (None, 64, 56, 56) 0 block_1a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_shortcut (Conv2D) (None, 256, 56, 56) 16384 activation_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_shortcut (BatchNorm (None, 256, 56, 56) 1024 block_1a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_1 (Add) (None, 256, 56, 56) 0 block_1a_bn_3[0][0] \n", " block_1a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu (Activation) (None, 256, 56, 56) 0 add_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_1 (Conv2D) (None, 64, 56, 56) 16384 block_1a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu_1 (Activation) (None, 64, 56, 56) 0 block_1b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu_2 (Activation) (None, 64, 56, 56) 0 block_1b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_2 (Add) (None, 256, 56, 56) 0 block_1b_bn_3[0][0] \n", " block_1a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu (Activation) (None, 256, 56, 56) 0 add_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_1 (Conv2D) (None, 64, 56, 56) 16384 block_1b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu_1 (Activation) (None, 64, 56, 56) 0 block_1c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu_2 (Activation) (None, 64, 56, 56) 0 block_1c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_3 (Add) (None, 256, 56, 56) 0 block_1c_bn_3[0][0] \n", " block_1b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu (Activation) (None, 256, 56, 56) 0 add_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_1 (Conv2D) (None, 128, 28, 28) 32768 block_1c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu_1 (Activation) (None, 128, 28, 28) 0 block_2a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu_2 (Activation) (None, 128, 28, 28) 0 block_2a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_shortcut (Conv2D) (None, 512, 28, 28) 131072 block_1c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_shortcut (BatchNorm (None, 512, 28, 28) 2048 block_2a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_4 (Add) (None, 512, 28, 28) 0 block_2a_bn_3[0][0] \n", " block_2a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu (Activation) (None, 512, 28, 28) 0 add_4[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu_1 (Activation) (None, 128, 28, 28) 0 block_2b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu_2 (Activation) (None, 128, 28, 28) 0 block_2b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_5 (Add) (None, 512, 28, 28) 0 block_2b_bn_3[0][0] \n", " block_2a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu (Activation) (None, 512, 28, 28) 0 add_5[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu_1 (Activation) (None, 128, 28, 28) 0 block_2c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu_2 (Activation) (None, 128, 28, 28) 0 block_2c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_6 (Add) (None, 512, 28, 28) 0 block_2c_bn_3[0][0] \n", " block_2b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu (Activation) (None, 512, 28, 28) 0 add_6[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2d_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu_1 (Activation) (None, 128, 28, 28) 0 block_2d_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2d_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2d_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu_2 (Activation) (None, 128, 28, 28) 0 block_2d_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2d_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2d_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_7 (Add) (None, 512, 28, 28) 0 block_2d_bn_3[0][0] \n", " block_2c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu (Activation) (None, 512, 28, 28) 0 add_7[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_1 (Conv2D) (None, 256, 14, 14) 131072 block_2d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu_1 (Activation) (None, 256, 14, 14) 0 block_3a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu_2 (Activation) (None, 256, 14, 14) 0 block_3a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_shortcut (Conv2D) (None, 1024, 14, 14) 524288 block_2d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_shortcut (BatchNorm (None, 1024, 14, 14) 4096 block_3a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_8 (Add) (None, 1024, 14, 14) 0 block_3a_bn_3[0][0] \n", " block_3a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu (Activation) (None, 1024, 14, 14) 0 add_8[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu_1 (Activation) (None, 256, 14, 14) 0 block_3b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu_2 (Activation) (None, 256, 14, 14) 0 block_3b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_9 (Add) (None, 1024, 14, 14) 0 block_3b_bn_3[0][0] \n", " block_3a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu (Activation) (None, 1024, 14, 14) 0 add_9[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu_1 (Activation) (None, 256, 14, 14) 0 block_3c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu_2 (Activation) (None, 256, 14, 14) 0 block_3c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_10 (Add) (None, 1024, 14, 14) 0 block_3c_bn_3[0][0] \n", " block_3b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu (Activation) (None, 1024, 14, 14) 0 add_10[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3d_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu_1 (Activation) (None, 256, 14, 14) 0 block_3d_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3d_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3d_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu_2 (Activation) (None, 256, 14, 14) 0 block_3d_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3d_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3d_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_11 (Add) (None, 1024, 14, 14) 0 block_3d_bn_3[0][0] \n", " block_3c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu (Activation) (None, 1024, 14, 14) 0 add_11[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3e_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu_1 (Activation) (None, 256, 14, 14) 0 block_3e_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3e_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3e_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu_2 (Activation) (None, 256, 14, 14) 0 block_3e_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3e_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3e_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_12 (Add) (None, 1024, 14, 14) 0 block_3e_bn_3[0][0] \n", " block_3d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu (Activation) (None, 1024, 14, 14) 0 add_12[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3e_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3f_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu_1 (Activation) (None, 256, 14, 14) 0 block_3f_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3f_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3f_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu_2 (Activation) (None, 256, 14, 14) 0 block_3f_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3f_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3f_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_13 (Add) (None, 1024, 14, 14) 0 block_3f_bn_3[0][0] \n", " block_3e_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu (Activation) (None, 1024, 14, 14) 0 add_13[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_1 (Conv2D) (None, 512, 14, 14) 524288 block_3f_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu_1 (Activation) (None, 512, 14, 14) 0 block_4a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu_2 (Activation) (None, 512, 14, 14) 0 block_4a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_shortcut (Conv2D) (None, 2048, 14, 14) 2097152 block_3f_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_shortcut (BatchNorm (None, 2048, 14, 14) 8192 block_4a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_14 (Add) (None, 2048, 14, 14) 0 block_4a_bn_3[0][0] \n", " block_4a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu (Activation) (None, 2048, 14, 14) 0 add_14[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_1 (Conv2D) (None, 512, 14, 14) 1048576 block_4a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu_1 (Activation) (None, 512, 14, 14) 0 block_4b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu_2 (Activation) (None, 512, 14, 14) 0 block_4b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_15 (Add) (None, 2048, 14, 14) 0 block_4b_bn_3[0][0] \n", " block_4a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu (Activation) (None, 2048, 14, 14) 0 add_15[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_1 (Conv2D) (None, 512, 14, 14) 1048576 block_4b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu_1 (Activation) (None, 512, 14, 14) 0 block_4c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu_2 (Activation) (None, 512, 14, 14) 0 block_4c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_16 (Add) (None, 2048, 14, 14) 0 block_4c_bn_3[0][0] \n", " block_4b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu (Activation) (None, 2048, 14, 14) 0 add_16[0][0] \n", "__________________________________________________________________________________________________\n", "avg_pool (AveragePooling2D) (None, 2048, 1, 1) 0 block_4c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "flatten (Flatten) (None, 2048) 0 avg_pool[0][0] \n", "__________________________________________________________________________________________________\n", "predictions (Dense) (None, 2) 4098 flatten[0][0] \n", "==================================================================================================\n", "Total params: 23,565,250\n", "Trainable params: 23,502,722\n", "Non-trainable params: 62,528\n", "__________________________________________________________________________________________________\n", "INFO: Processing /workspace/tao-experiments/data/val/notdefect...\n", "INFO: Inference complete. Result is saved at /workspace/tao-experiments/data/val/notdefect/result.csv\n", "Telemetry data couldn't be sent, but the command ran successfully.\n", "[WARNING]: \n", "Execution status: PASS\n", "2026-02-02 18:08:41,020 [TAO Toolkit] [INFO] nvidia_tao_cli.components.docker_handler.docker_handler 363: Stopping container.\n" ] } ], "source": [ "!tao model classification_tf1 inference -m $TAO_EXPERIMENT_DIR/resnet50_aug/weights/resnet_010.hdf5 \\\n", " -d $TAO_DATA_DIR/val/notdefect \\\n", " -e $TAO_SPECS_DIR/resnet50/combined_config_aug.txt \\\n", " -cm $TAO_EXPERIMENT_DIR/resnet50/classmap.json \\\n", " -k $KEY" ] }, { "cell_type": "code", "execution_count": 44, "id": "72eead30-22b2-4c9a-ab08-fe32dcf39f4d", "metadata": { "scrolled": true, "tags": [] }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "2026-02-02 18:08:41,661 [TAO Toolkit] [INFO] root 160: Registry: ['nvcr.io']\n", "2026-02-02 18:08:41,759 [TAO Toolkit] [INFO] nvidia_tao_cli.components.instance_handler.local_instance 360: Running command in container: nvcr.io/nvidia/tao/tao-toolkit:5.0.0-tf1.15.5\n", "2026-02-02 18:08:41,769 [TAO Toolkit] [INFO] nvidia_tao_cli.components.docker_handler.docker_handler 301: Printing tty value True\n", "Using TensorFlow backend.\n", "2026-02-02 18:08:42.583043: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 18:08:42,633 [TAO Toolkit] [WARNING] tensorflow 40: Deprecation warnings have been disabled. Set TF_ENABLE_DEPRECATION_WARNINGS=1 to re-enable them.\n", "2026-02-02 18:08:43,687 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use sklearn by default. This improves performance in some cases. To enable sklearn export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 18:08:43,726 [TAO Toolkit] [WARNING] tensorflow 42: TensorFlow will not use Dask by default. This improves performance in some cases. To enable Dask export the environment variable TF_ALLOW_IOLIBS=1.\n", "2026-02-02 18:08:43,730 [TAO Toolkit] [WARNING] tensorflow 43: TensorFlow will not use Pandas by default. This improves performance in some cases. To enable Pandas export the environment variable TF_ALLOW_IOLIBS=1.\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:150: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_hue(img, max_delta=10.0):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:173: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_saturation(img, max_shift):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:183: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_contrast(img, center, max_contrast_scale):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:192: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_shift(x_img, shift_stddev):\n", "2026-02-02 18:08:45.576545: I tensorflow/core/platform/profile_utils/cpu_utils.cc:109] CPU Frequency: 2499995000 Hz\n", "2026-02-02 18:08:45.576871: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x89611f0 initialized for platform Host (this does not guarantee that XLA will be used). Devices:\n", "2026-02-02 18:08:45.576909: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version\n", "2026-02-02 18:08:45.578438: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcuda.so.1\n", "2026-02-02 18:08:45.767218: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:08:45.769077: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x845a5f0 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:\n", "2026-02-02 18:08:45.769109: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Tesla T4, Compute Capability 7.5\n", "2026-02-02 18:08:45.769432: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:08:45.771099: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1674] Found device 0 with properties: \n", "name: Tesla T4 major: 7 minor: 5 memoryClockRate(GHz): 1.59\n", "pciBusID: 0000:00:1e.0\n", "2026-02-02 18:08:45.771162: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 18:08:45.771298: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcublas.so.12\n", "2026-02-02 18:08:45.773723: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcufft.so.11\n", "2026-02-02 18:08:45.773813: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcurand.so.10\n", "2026-02-02 18:08:45.777511: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcusolver.so.11\n", "2026-02-02 18:08:45.778693: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcusparse.so.12\n", "2026-02-02 18:08:45.778768: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudnn.so.8\n", "2026-02-02 18:08:45.778899: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:08:45.780616: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:08:45.782230: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1802] Adding visible gpu devices: 0\n", "2026-02-02 18:08:45.782284: I tensorflow/stream_executor/platform/default/dso_loader.cc:50] Successfully opened dynamic library libcudart.so.12\n", "2026-02-02 18:08:45.790783: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1214] Device interconnect StreamExecutor with strength 1 edge matrix:\n", "2026-02-02 18:08:45.790815: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1220] 0 \n", "2026-02-02 18:08:45.790829: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1233] 0: N \n", "2026-02-02 18:08:45.791040: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:08:45.792766: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1082] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n", "2026-02-02 18:08:45.794401: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1359] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 13332 MB memory) -> physical GPU (device: 0, name: Tesla T4, pci bus id: 0000:00:1e.0, compute capability: 7.5)\n", "Using TensorFlow backend.\n", "WARNING: Deprecation warnings have been disabled. Set TF_ENABLE_DEPRECATION_WARNINGS=1 to re-enable them.\n", "WARNING: TensorFlow will not use sklearn by default. This improves performance in some cases. To enable sklearn export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING: TensorFlow will not use Dask by default. This improves performance in some cases. To enable Dask export the environment variable TF_ALLOW_IOLIBS=1.\n", "WARNING: TensorFlow will not use Pandas by default. This improves performance in some cases. To enable Pandas export the environment variable TF_ALLOW_IOLIBS=1.\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:150: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_hue(img, max_delta=10.0):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:173: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_saturation(img, max_shift):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:183: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_contrast(img, center, max_contrast_scale):\n", "/usr/local/lib/python3.8/dist-packages/nvidia_tao_tf1/cv/makenet/utils/helper.py:192: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details.\n", " def random_shift(x_img, shift_stddev):\n", "INFO: Loading experiment spec at /workspace/tao-experiments/spec_files/resnet50/combined_config_aug.txt.\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:245: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:1834: The name tf.nn.fused_batch_norm is deprecated. Please use tf.compat.v1.nn.fused_batch_norm instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/third_party/keras/tensorflow_backend.py:199: The name tf.nn.avg_pool is deprecated. Please use tf.nn.avg_pool2d instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:174: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:181: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:186: The name tf.Session is deprecated. Please use tf.compat.v1.Session instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:190: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:199: The name tf.is_variable_initialized is deprecated. Please use tf.compat.v1.is_variable_initialized instead.\n", "\n", "WARNING: From /usr/local/lib/python3.8/dist-packages/keras/backend/tensorflow_backend.py:206: The name tf.variables_initializer is deprecated. Please use tf.compat.v1.variables_initializer instead.\n", "\n", "__________________________________________________________________________________________________\n", "Layer (type) Output Shape Param # Connected to \n", "==================================================================================================\n", "input_1 (InputLayer) (None, 3, 224, 224) 0 \n", "__________________________________________________________________________________________________\n", "conv1 (Conv2D) (None, 64, 112, 112) 9408 input_1[0][0] \n", "__________________________________________________________________________________________________\n", "bn_conv1 (BatchNormalization) (None, 64, 112, 112) 256 conv1[0][0] \n", "__________________________________________________________________________________________________\n", "activation_1 (Activation) (None, 64, 112, 112) 0 bn_conv1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_1 (Conv2D) (None, 64, 56, 56) 4096 activation_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu_1 (Activation) (None, 64, 56, 56) 0 block_1a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu_2 (Activation) (None, 64, 56, 56) 0 block_1a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_conv_shortcut (Conv2D) (None, 256, 56, 56) 16384 activation_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_bn_shortcut (BatchNorm (None, 256, 56, 56) 1024 block_1a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_1 (Add) (None, 256, 56, 56) 0 block_1a_bn_3[0][0] \n", " block_1a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_1a_relu (Activation) (None, 256, 56, 56) 0 add_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_1 (Conv2D) (None, 64, 56, 56) 16384 block_1a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu_1 (Activation) (None, 64, 56, 56) 0 block_1b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu_2 (Activation) (None, 64, 56, 56) 0 block_1b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_2 (Add) (None, 256, 56, 56) 0 block_1b_bn_3[0][0] \n", " block_1a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1b_relu (Activation) (None, 256, 56, 56) 0 add_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_1 (Conv2D) (None, 64, 56, 56) 16384 block_1b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_1 (BatchNormalizati (None, 64, 56, 56) 256 block_1c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu_1 (Activation) (None, 64, 56, 56) 0 block_1c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_2 (Conv2D) (None, 64, 56, 56) 36864 block_1c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_2 (BatchNormalizati (None, 64, 56, 56) 256 block_1c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu_2 (Activation) (None, 64, 56, 56) 0 block_1c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_conv_3 (Conv2D) (None, 256, 56, 56) 16384 block_1c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_bn_3 (BatchNormalizati (None, 256, 56, 56) 1024 block_1c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_3 (Add) (None, 256, 56, 56) 0 block_1c_bn_3[0][0] \n", " block_1b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_1c_relu (Activation) (None, 256, 56, 56) 0 add_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_1 (Conv2D) (None, 128, 28, 28) 32768 block_1c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu_1 (Activation) (None, 128, 28, 28) 0 block_2a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu_2 (Activation) (None, 128, 28, 28) 0 block_2a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_conv_shortcut (Conv2D) (None, 512, 28, 28) 131072 block_1c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_bn_shortcut (BatchNorm (None, 512, 28, 28) 2048 block_2a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_4 (Add) (None, 512, 28, 28) 0 block_2a_bn_3[0][0] \n", " block_2a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_2a_relu (Activation) (None, 512, 28, 28) 0 add_4[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu_1 (Activation) (None, 128, 28, 28) 0 block_2b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu_2 (Activation) (None, 128, 28, 28) 0 block_2b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_5 (Add) (None, 512, 28, 28) 0 block_2b_bn_3[0][0] \n", " block_2a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2b_relu (Activation) (None, 512, 28, 28) 0 add_5[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu_1 (Activation) (None, 128, 28, 28) 0 block_2c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu_2 (Activation) (None, 128, 28, 28) 0 block_2c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_6 (Add) (None, 512, 28, 28) 0 block_2c_bn_3[0][0] \n", " block_2b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2c_relu (Activation) (None, 512, 28, 28) 0 add_6[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_1 (Conv2D) (None, 128, 28, 28) 65536 block_2c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_1 (BatchNormalizati (None, 128, 28, 28) 512 block_2d_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu_1 (Activation) (None, 128, 28, 28) 0 block_2d_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_2 (Conv2D) (None, 128, 28, 28) 147456 block_2d_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_2 (BatchNormalizati (None, 128, 28, 28) 512 block_2d_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu_2 (Activation) (None, 128, 28, 28) 0 block_2d_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_conv_3 (Conv2D) (None, 512, 28, 28) 65536 block_2d_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_bn_3 (BatchNormalizati (None, 512, 28, 28) 2048 block_2d_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_7 (Add) (None, 512, 28, 28) 0 block_2d_bn_3[0][0] \n", " block_2c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_2d_relu (Activation) (None, 512, 28, 28) 0 add_7[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_1 (Conv2D) (None, 256, 14, 14) 131072 block_2d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu_1 (Activation) (None, 256, 14, 14) 0 block_3a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu_2 (Activation) (None, 256, 14, 14) 0 block_3a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_conv_shortcut (Conv2D) (None, 1024, 14, 14) 524288 block_2d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_bn_shortcut (BatchNorm (None, 1024, 14, 14) 4096 block_3a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_8 (Add) (None, 1024, 14, 14) 0 block_3a_bn_3[0][0] \n", " block_3a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_3a_relu (Activation) (None, 1024, 14, 14) 0 add_8[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu_1 (Activation) (None, 256, 14, 14) 0 block_3b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu_2 (Activation) (None, 256, 14, 14) 0 block_3b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_9 (Add) (None, 1024, 14, 14) 0 block_3b_bn_3[0][0] \n", " block_3a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3b_relu (Activation) (None, 1024, 14, 14) 0 add_9[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu_1 (Activation) (None, 256, 14, 14) 0 block_3c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu_2 (Activation) (None, 256, 14, 14) 0 block_3c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_10 (Add) (None, 1024, 14, 14) 0 block_3c_bn_3[0][0] \n", " block_3b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3c_relu (Activation) (None, 1024, 14, 14) 0 add_10[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3d_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu_1 (Activation) (None, 256, 14, 14) 0 block_3d_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3d_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3d_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu_2 (Activation) (None, 256, 14, 14) 0 block_3d_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3d_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3d_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_11 (Add) (None, 1024, 14, 14) 0 block_3d_bn_3[0][0] \n", " block_3c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3d_relu (Activation) (None, 1024, 14, 14) 0 add_11[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3e_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu_1 (Activation) (None, 256, 14, 14) 0 block_3e_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3e_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3e_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu_2 (Activation) (None, 256, 14, 14) 0 block_3e_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3e_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3e_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_12 (Add) (None, 1024, 14, 14) 0 block_3e_bn_3[0][0] \n", " block_3d_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3e_relu (Activation) (None, 1024, 14, 14) 0 add_12[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_1 (Conv2D) (None, 256, 14, 14) 262144 block_3e_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_1 (BatchNormalizati (None, 256, 14, 14) 1024 block_3f_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu_1 (Activation) (None, 256, 14, 14) 0 block_3f_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_2 (Conv2D) (None, 256, 14, 14) 589824 block_3f_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_2 (BatchNormalizati (None, 256, 14, 14) 1024 block_3f_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu_2 (Activation) (None, 256, 14, 14) 0 block_3f_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_conv_3 (Conv2D) (None, 1024, 14, 14) 262144 block_3f_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_bn_3 (BatchNormalizati (None, 1024, 14, 14) 4096 block_3f_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_13 (Add) (None, 1024, 14, 14) 0 block_3f_bn_3[0][0] \n", " block_3e_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_3f_relu (Activation) (None, 1024, 14, 14) 0 add_13[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_1 (Conv2D) (None, 512, 14, 14) 524288 block_3f_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4a_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu_1 (Activation) (None, 512, 14, 14) 0 block_4a_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4a_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4a_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu_2 (Activation) (None, 512, 14, 14) 0 block_4a_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4a_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_conv_shortcut (Conv2D) (None, 2048, 14, 14) 2097152 block_3f_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4a_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_bn_shortcut (BatchNorm (None, 2048, 14, 14) 8192 block_4a_conv_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "add_14 (Add) (None, 2048, 14, 14) 0 block_4a_bn_3[0][0] \n", " block_4a_bn_shortcut[0][0] \n", "__________________________________________________________________________________________________\n", "block_4a_relu (Activation) (None, 2048, 14, 14) 0 add_14[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_1 (Conv2D) (None, 512, 14, 14) 1048576 block_4a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4b_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu_1 (Activation) (None, 512, 14, 14) 0 block_4b_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4b_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4b_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu_2 (Activation) (None, 512, 14, 14) 0 block_4b_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4b_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4b_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_15 (Add) (None, 2048, 14, 14) 0 block_4b_bn_3[0][0] \n", " block_4a_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4b_relu (Activation) (None, 2048, 14, 14) 0 add_15[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_1 (Conv2D) (None, 512, 14, 14) 1048576 block_4b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_1 (BatchNormalizati (None, 512, 14, 14) 2048 block_4c_conv_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu_1 (Activation) (None, 512, 14, 14) 0 block_4c_bn_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_2 (Conv2D) (None, 512, 14, 14) 2359296 block_4c_relu_1[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_2 (BatchNormalizati (None, 512, 14, 14) 2048 block_4c_conv_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu_2 (Activation) (None, 512, 14, 14) 0 block_4c_bn_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_conv_3 (Conv2D) (None, 2048, 14, 14) 1048576 block_4c_relu_2[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_bn_3 (BatchNormalizati (None, 2048, 14, 14) 8192 block_4c_conv_3[0][0] \n", "__________________________________________________________________________________________________\n", "add_16 (Add) (None, 2048, 14, 14) 0 block_4c_bn_3[0][0] \n", " block_4b_relu[0][0] \n", "__________________________________________________________________________________________________\n", "block_4c_relu (Activation) (None, 2048, 14, 14) 0 add_16[0][0] \n", "__________________________________________________________________________________________________\n", "avg_pool (AveragePooling2D) (None, 2048, 1, 1) 0 block_4c_relu[0][0] \n", "__________________________________________________________________________________________________\n", "flatten (Flatten) (None, 2048) 0 avg_pool[0][0] \n", "__________________________________________________________________________________________________\n", "predictions (Dense) (None, 2) 4098 flatten[0][0] \n", "==================================================================================================\n", "Total params: 23,565,250\n", "Trainable params: 23,502,722\n", "Non-trainable params: 62,528\n", "__________________________________________________________________________________________________\n", "INFO: Processing /workspace/tao-experiments/data/val/defect...\n", "INFO: Inference complete. Result is saved at /workspace/tao-experiments/data/val/defect/result.csv\n", "Telemetry data couldn't be sent, but the command ran successfully.\n", "[WARNING]: \n", "Execution status: PASS\n", "2026-02-02 18:09:21,390 [TAO Toolkit] [INFO] nvidia_tao_cli.components.docker_handler.docker_handler 363: Stopping container.\n" ] } ], "source": [ "!tao model classification_tf1 inference -m $TAO_EXPERIMENT_DIR/resnet50_aug/weights/resnet_010.hdf5 \\\n", " -d $TAO_DATA_DIR/val/defect \\\n", " -e $TAO_SPECS_DIR/resnet50/combined_config_aug.txt \\\n", " -cm $TAO_EXPERIMENT_DIR/resnet50/classmap.json \\\n", " -k $KEY" ] }, { "cell_type": "markdown", "id": "015a06ad-4c34-4d20-bccd-ddec8af4e18d", "metadata": {}, "source": [ "When executing with `-d`, or directory mode, a `result.csv` file is created and stored in the directory you specify using `-d`. The `result.csv` has the file path in the first column and predicted labels in the second. " ] }, { "cell_type": "markdown", "id": "f66c94ef-d9ea-466c-8504-469f34ae3758", "metadata": {}, "source": [ "\n", "### Precision and Recall ###\n", "We can set a threshold on the probability score and classify only images above this threshold as true defects. In order to accomplish this, we'll have to understand two measures: `recall` and `precision`. The first measure is focused on identifying positive cases and is called __recall__. We define recall as the ability of the model to identify all true positive samples of the dataset. In mathematical terms, recall is the ratio of true positives over true positives plus false negatives. By other means, recall tells us, among all the test samples belonging to the output class, how many of them are identified correctly by the model. The next measure, is called __precision__ and is the ability of the model to identify the relevant samples only, and is defined as the ratio of true positives over true positives plus false positives. Selecting a proper threshold, usually stems from a good balance between the precision and recall values. A well-known measure that provides such a balance is __f1-score__, which is a harmonic mean of precision and recall. \n", "

" ] }, { "cell_type": "code", "execution_count": 45, "id": "f6e80025-20c5-4f97-b0b8-ce2df0bf49c6", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
image_pathpredictionprobabilitytrue_defect
0/workspace/tao-experiments/data/val/notdefect/...notdefect0.993725notdefect
1/workspace/tao-experiments/data/val/notdefect/...notdefect0.941579notdefect
2/workspace/tao-experiments/data/val/notdefect/...notdefect0.669868notdefect
3/workspace/tao-experiments/data/val/notdefect/...notdefect0.985907notdefect
4/workspace/tao-experiments/data/val/notdefect/...notdefect0.969943notdefect
\n", "
" ], "text/plain": [ " image_path prediction probability \\\n", "0 /workspace/tao-experiments/data/val/notdefect/... notdefect 0.993725 \n", "1 /workspace/tao-experiments/data/val/notdefect/... notdefect 0.941579 \n", "2 /workspace/tao-experiments/data/val/notdefect/... notdefect 0.669868 \n", "3 /workspace/tao-experiments/data/val/notdefect/... notdefect 0.985907 \n", "4 /workspace/tao-experiments/data/val/notdefect/... notdefect 0.969943 \n", "\n", " true_defect \n", "0 notdefect \n", "1 notdefect \n", "2 notdefect \n", "3 notdefect \n", "4 notdefect " ] }, "execution_count": 45, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# read the results csv files\n", "notdefect_df=pd.read_csv(os.path.join(os.environ['LOCAL_DATA_DIR'], 'val/notdefect/result.csv'), names=['image_path', 'prediction', 'probability'])\n", "notdefect_df['true_defect']='notdefect'\n", "defect_df=pd.read_csv(os.path.join(os.environ['LOCAL_DATA_DIR'], 'val/defect/result.csv'), names=['image_path', 'prediction', 'probability'])\n", "defect_df['true_defect']='defect'\n", "results_df=pd.concat([notdefect_df, defect_df])\n", "\n", "# preview the results dataframe\n", "results_df.head()" ] }, { "cell_type": "code", "execution_count": 46, "id": "77adb013-0323-4ea7-bdeb-4494098b4376", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
predictiondefectnotdefect
true_defect
defect129
notdefect0541
\n", "
" ], "text/plain": [ "prediction defect notdefect\n", "true_defect \n", "defect 1 29\n", "notdefect 0 541" ] }, "execution_count": 46, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# create confusion matrix\n", "confusion_df=pd.crosstab(results_df['true_defect'], results_df['prediction'])\n", "\n", "# preview the confusion matrix dataframe\n", "confusion_df.head()" ] }, { "cell_type": "code", "execution_count": 47, "id": "d6499681-6d6a-4892-95e6-51e210fd5e1a", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Recall is: 0.03\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# calculate the recall for defects\n", "try: \n", " actual_defect=confusion_df.sum(axis=1)['defect']\n", " recall=confusion_df.loc['defect', 'defect']/actual_defect\n", " print('Recall is: {}'.format(round(recall, 2)))\n", "except: \n", " print('Cannot calculate recall.')" ] }, { "cell_type": "code", "execution_count": 48, "id": "5e479694-2682-48af-a937-ad8e6ef154a5", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Precision is: 1.0\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# calculate the precision for defects\n", "try: \n", " predicted_defect=confusion_df['defect'].sum()\n", " precision=confusion_df.loc['defect', 'defect']/predicted_defect\n", " print('Precision is: {}'.format(round(precision, 2)))\n", "except: \n", " print('Cannot calculate precision.')" ] }, { "cell_type": "code", "execution_count": 49, "id": "412046e7-1ba6-4c58-a550-715d06c81fa5", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "F1-score is: 0.06\n" ] } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# calculate the f1-score for defects\n", "try: \n", " f1_score=2*precision*recall/(precision+recall)\n", " print('F1-score is: {}'.format(round(f1_score, 2)))\n", "except: \n", " print('Cannot calculate f1-score.')" ] }, { "cell_type": "code", "execution_count": 50, "id": "b6f845d1-6b4c-4064-bfe4-4ab455c3c1f6", "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "# plot bad predictions\n", "plt.figure(figsize=(15, 15))\n", "\n", "errors_df=results_df[results_df['prediction']!=results_df['true_defect']].reset_index(drop=True)\n", "\n", "num_cols=5\n", "num_rows=math.ceil(len(errors_df)/num_cols)\n", "\n", "for i, row in errors_df.iterrows(): \n", " image_path=row['image_path']\n", " image_path=image_path.replace(os.environ['TAO_DATA_DIR'], os.environ['LOCAL_DATA_DIR'])\n", " plt.subplot(num_rows, num_cols, i+1)\n", " plt.imshow(mpimg.imread(image_path))\n", " plt.title('Predicted: {} \\n Actual: {}'.format(row['prediction'], row['true_defect']))\n", " plt.axis('off')" ] }, { "cell_type": "markdown", "id": "4203951c-162e-49f8-9bc5-2ee579f45c09", "metadata": {}, "source": [ "\n", "### Adjusting the Threshold ###\n", "With imbalanced datasets, the Area Under the Curve (AUC) score calculated from the Receiver Operating Characteristic (ROC) is a very useful metric for identifying the optimum threshold for classification. AUC - ROC curve is a performance measurement for a classification problem at various thresholds settings where ROC is a probability curve and AUC represents degree or measure of separability. The ROC curve is plotted with TPR against the FPR where TPR is on y-axis and FPR is on the x-axis. TPR and FPR are defined as follows:\n", "* TPR = True Positives / All Positives\n", "* FPR = False Positives / All negatives\n", "\n", "With ROC, we can see the model's performance at various classification thresholds: \n", "

" ] }, { "cell_type": "code", "execution_count": 51, "id": "044a6a2e-dc62-4096-ba90-a810e0171c5d", "metadata": {}, "outputs": [], "source": [ "# DO NOT CHANGE THIS CELL\n", "def get_single_tpr_fpr(df):\n", " '''\n", " calculate true positive rate and false positive rate\n", " \n", " param: \n", " df: the dataframe should have 'true_defect' and 'prediction' as its labels\n", " return: \n", " a list containing tpr and fpr\n", " '''\n", " \n", " tp=((df['true_defect'] == 'defect' ) & (df['prediction'] == 'defect')).sum()\n", " fp=((df['true_defect'] == 'notdefect' ) & (df['prediction'] == 'defect')).sum()\n", " tn=((df['true_defect'] == 'notdefect' ) & (df['prediction'] == 'notdefect')).sum()\n", " fn=((df['true_defect'] == 'defect' ) & (df['prediction'] == 'notdefect')).sum()\n", "\n", " tpr=tp / (tp + fn )\n", " fpr=fp / (fp + tn)\n", "\n", " return [tpr, fpr]" ] }, { "cell_type": "code", "execution_count": 52, "id": "26d80651-c3cf-4ad6-b741-1e44990578ce", "metadata": {}, "outputs": [], "source": [ "# DO NOT CHANGE THIS CELL\n", "# while computing AUC score you need to calculate \"TP,\"FP\" at every threshold by using actual \"true_positive\" and predicted \"prediction\".\n", "def calculate_all_thresholds_tpr_fpr_arr(df):\n", " '''\n", " calculate true positive rates and false positive rates at various thresholds\n", " \n", " param: \n", " df: the original dataframe, which should have a 'probability' label\n", " return: \n", " two arrays: tpr_arr_for_all_thresholds, fpr_arr_for_all_thresholds\n", " '''\n", " tpr_arr_for_all_thresholds = []\n", " fpr_arr_for_all_thresholds = []\n", "\n", " sorted_df=df.sort_values(by=['probability'], ascending=False)\n", " sorted_df.loc[sorted_df['prediction']=='notdefect', 'probability']=1-sorted_df.loc[sorted_df['prediction']=='notdefect', 'probability']\n", "\n", " unique_probability_thresholds = sorted_df['probability'].unique()\n", "\n", " for threshold in unique_probability_thresholds:\n", " sorted_df['prediction'] = np.where(sorted_df['probability'] >= threshold, 'defect', 'notdefect')\n", " tpr_fpr_arr = get_single_tpr_fpr(sorted_df)\n", " tpr_arr_for_all_thresholds.append(tpr_fpr_arr[0])\n", " fpr_arr_for_all_thresholds.append(tpr_fpr_arr[1])\n", "\n", " return tpr_arr_for_all_thresholds, fpr_arr_for_all_thresholds" ] }, { "cell_type": "code", "execution_count": 53, "id": "fbbdf9c9-fa0c-4dfc-a463-faf3179bbb5c", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 53, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "fig=plt.figure(figsize=(5, 5))\n", "all_tpr, all_fpr=calculate_all_thresholds_tpr_fpr_arr(results_df)\n", "plt.scatter(all_fpr, all_tpr, s=1)" ] }, { "cell_type": "code", "execution_count": 54, "id": "f9998a18-d130-4d20-b210-709495159c6b", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "array([[,\n", " ]], dtype=object)" ] }, "execution_count": 54, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# DO NOT CHANGE THIS CELL\n", "results_df.pivot_table(columns='prediction', index='image_path', values='probability').hist(stacked=True)" ] }, { "cell_type": "markdown", "id": "01280ca6-e986-4fe6-9ec5-366e8cf6df8a", "metadata": {}, "source": [ "**Well Done!** When you're finished, please complete the assessment before moving onto the next lab. " ] }, { "cell_type": "markdown", "id": "5e715988-e892-4565-bcb9-85c289eefd2d", "metadata": {}, "source": [ " \"Header\" " ] } ], "metadata": { "kernelspec": { "display_name": "Python 3 (ipykernel)", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.10.12" } }, "nbformat": 4, "nbformat_minor": 5 }