This resource is using open-source code maintained in github (see the quick-start-guide section) and available for download from NGC

The nnU-Net (\"no-new-Net\") refers to a robust and self-adapting framework for U-Net based medical image segmentation. This repository contains a nnU-Net implementation as described in the paper: nnU-Net: Self-adapting Framework for U-Net-Based Medical Image Segmentation.

The differences between this nnU-net and the original model are:

Dynamic selection of patch size is not supported, and it has to be set in data_preprocessing/configs.py file.
Cascaded U-Net is not supported.
The following data augmentations are not used: rotation, simulation of low resolution, gamma augmentation.

This model is trained with mixed precision using Tensor Cores on Volta, Turing, and the NVIDIA Ampere GPU architectures. Therefore, researchers can get results 2x faster than training without Tensor Cores, while experiencing the benefits of mixed precision training. This model is tested against each NGC monthly container release to ensure consistent accuracy and performance over time.

Model architecture

The nnU-Net allows training two types of networks: 2D U-Net and 3D U-Net to perform semantic segmentation of 2D or 3D images, with high accuracy and performance.

The following figure shows the architecture of the 3D U-Net model and its different components. U-Net is composed of a contractive and an expanding path, that aims at building a bottleneck in its centermost part through a combination of convolution, instance norm, and leaky ReLU operations. After this bottleneck, the image is reconstructed through a combination of convolutions and upsampling. Skip connections are added with the goal of helping the backward flow of gradients to improve the training.

Figure 1: The 3D U-Net architecture

Default configuration

All convolution blocks in U-Net in both encoder and decoder are using two convolution layers followed by instance normalization and a leaky ReLU nonlinearity. For downsampling, we are using stride convolution whereas transposed convolution is used for upsampling.

All models were trained with the Adam optimizer. For loss function we use the average of cross-entropy and dice coefficient.

Used data augmentation: crop with oversampling the foreground class, mirroring, zoom, Gaussian noise, Gaussian blur, brightness.

Feature support matrix

The following features are supported by this model:

\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

tr:nth-child(odd)]:bg-surface-sunken\">

Feature	nnUNet
DALI	Yes
Automatic mixed precision (AMP)	Yes
Horovod Multi-GPU (NCCL)	Yes
XLA	Yes

Features

DALI\nNVIDIA Data Loading Library (DALI) is a collection of optimized building blocks, and an execution engine, to speed up the pre-processing of the input data for deep learning applications. DALI provides both the performance and the flexibility for accelerating different data pipelines as a single library. This single library can then be integrated into different deep learning training and inference applications. For details, refer to example sources in this repository or refer to the DALI documentation.

Automatic Mixed Precision (AMP)\nComputation graphs can be modified by TensorFlow during runtime to support mixed precision training, which allows using FP16 training with FP32 master weights. A detailed explanation of mixed precision can be found in the next section.

Multi-GPU training with Horovod\nHorovod is a distributed training framework for TensorFlow, Keras, PyTorch, and MXNet. The goal of Horovod is to make distributed deep learning fast and easy to use. For more information about how to get started with Horovod, refer to the Horovod: Official repository.\nOur model uses Horovod to implement efficient multi-GPU training with NCCL. For details, refer to example scripts in this repository or refer to the TensorFlow tutorial.

XLA\nXLA (Accelerated Linear Algebra) is a compiler that can speed up TensorFlow networks by model-specific optimizations i.e. fusing many GPU operations together.\nOperations fused into a single GPU kernel do not have to use extra memory to store intermediate values by keeping them in GPU registers, thus reducing memory operations and improving performance. For details refer to the TensorFlow documentation.

Mixed precision training

Mixed precision is the combined use of different numerical precisions in a computational method. Mixed precision training offers significant computational speedup by performing operations in the half-precision format while storing minimal information in single-precision to keep as much information as possible in critical parts of the network. Since the introduction of Tensor Cores in Volta, and following with both the Turing and Ampere architectures, significant training speedups are experienced by switching to mixed precision -- up to 3x speedup on the most intense model architectures. Using mixed precision training requires two steps:

Porting the model to use the FP16 data type where appropriate.
Adding loss scaling to preserve small gradient values.

This can now be achieved using Automatic Mixed Precision (AMP) for TensorFlow to enable the full mixed precision methodology in your existing TensorFlow model code. AMP enables mixed precision training on NVIDIA Volta, NVIDIA Turing, and NVIDIA Ampere GPU architectures automatically. The TensorFlow framework code makes all necessary model changes internally.

In TF-AMP, the computational graph is optimized to use as few casts as necessary and maximize the use of FP16, and the loss scaling is automatically applied inside of supported optimizers. AMP can be configured to work with the existing tf.contrib loss scaling manager by disabling the AMP scaling with a single environment variable to perform only the automatic mixed-precision optimization. It accomplishes this by automatically rewriting all computation graphs with the necessary operations to enable mixed precision training and automatic loss scaling.

For information about:

How to train using mixed precision, see the Mixed Precision Training paper and Training With Mixed Precision documentation.
Techniques used for mixed precision training, see the Mixed-Precision Training of Deep Neural Networks blog.

Enabling mixed precision

Mixed precision is enabled in TensorFlow by using the Automatic Mixed Precision (TF-AMP) extension which casts variables to half-precision upon retrieval, while storing variables in single-precision format. Furthermore, to preserve small gradient magnitudes in backpropagation, a loss scaling step must be included when applying gradients. In TensorFlow, loss scaling can be applied statically by using simple multiplication of loss by a constant value or automatically, by TF-AMP. Automatic mixed precision makes all the adjustments internally in TensorFlow, providing two benefits over manual operations. First, programmers need not modify network model code, reducing development and maintenance efforts. Second, using AMP maintains forward and backward compatibility with all the APIs for defining and running TensorFlow models.

Example nnU-Net scripts for training, inference, and benchmarking from the scripts/ directory enable mixed precision if the --amp command line flag is used.

Internally, mixed precision is enabled by setting keras.mixed_precision policy to mixed_float16. Additionally, our custom training loop uses a LossScaleOptimizer wrapper for the optimizer. For more information see the Mixed precision guide.

TF32

TensorFloat-32 (TF32) is the new math mode in NVIDIA A100 GPUs for handling the matrix math, also called tensor operations. TF32 running on Tensor Cores in A100 GPUs can provide up to 10x speedups compared to single-precision floating-point math (FP32) on NVIDIA Volta GPUs.

TF32 Tensor Cores can speed up networks using FP32, typically with no loss of accuracy. It is more robust than FP16 for models which require a high dynamic range for weights or activations.

For more information, refer to the TensorFloat-32 in the A100 GPU Accelerates AI Training, HPC up to 20x blog post.

TF32 is supported in the NVIDIA Ampere GPU architecture and is enabled by default.

Glossary

Deep supervision\nDeep supervision is a technique that adds auxiliary loss outputs to the U-Net decoder layers. For nnU-Net, we add auxiliary losses to the three latest decoder levels. The final loss is a weighted average of the obtained loss values. Deep supervision can be enabled by adding the --deep-supervision flag.

Test time augmentation\nTest time augmentation is an inference technique that averages predictions from augmented images with its prediction. As a result, predictions are more accurate, but with the cost of a slower inference process. For nnU-Net, we use all possible flip combinations for image augmenting. Test time augmentation can be enabled by adding the --tta flag to the training or inference script invocation.

Sliding window inference\nDuring inference, this method replaces an arbitrary resolution input image with a batch of overlapping windows, that cover the whole input. After passing this batch through the network a prediction with the original resolution is reassembled. Predicted values inside overlapped regions are obtained from a weighted average.\nOverlap ratio and weights for the average (i.e. blending mode) can be adjusted with the --overlap and --blend-mode options.

","resourceVersions",[236,237,238],"resourceVersionsTotal","hasMoreVersions","versionMetadata",{"_215":29,"_216":-5,"_53":120,"_117":-5,"_121":217,"_123":-5,"_218":134,"_128":18,"_125":35,"_219":-5,"_124":35,"_220":221,"_222":92,"_223":224,"_225":92,"_226":227,"_228":92,"_229":230,"_231":92,"_232":233,"_234":92,"_126":235},"versionNotFound","isLoggedIn","latestVersionPath","/orgs/nvidia/dle/resources/nnunet_tf2/-","imageReference","cliDownloadCommand","ngc registry resource download-version nvidia/dle/nnunet_tf2:22.11.0","helmFetchCommand","viewAllVersionsPath","/orgs/nvidia/dle/resources/nnunet_tf2/22.11.0/version-history","accessCheckResult",{"_214":92},"hasRelatedCollections","context",{"_210":211,"_212":213},1789162843036,[],[],[204,205,206,207,208,209],{"_49":64,"_51":75,"_53":66,"_55":67,"_57":76,"_59":60},{"_49":64,"_51":71,"_53":66,"_55":67,"_57":72,"_59":60},{"_49":64,"_51":69,"_53":66,"_55":67,"_57":70,"_59":60},{"_49":64,"_51":65,"_53":66,"_55":67,"_57":68,"_59":60},{"_49":61,"_51":62,"_53":153,"_55":-5,"_57":-5,"_59":60},{"_49":50,"_51":52,"_53":54,"_55":56,"_57":58,"_59":60},"orgName","nvidia","teamName","dle","allowed","versionId","description",["D",1670257650857],"totalSizeInBytes","malwareScanStatus","releaseNotes","### Changelog\n\nNovember 2022\n- Container update to 22.11\n- Use channel last layout for convolution with XLA\n- Add support for GPU binding\n\nMay 2022\n- Initial release","hasReleaseNotes","setup","The following section lists the requirements that you need to meet in order to start training the nnU-Net model.\n\n### Requirements\n\nThis repository contains Dockerfile which extends the TensorFlow 2 NGC container and encapsulates some dependencies. Aside from these dependencies, ensure you have the following components:\n- [NVIDIA Docker](https://github.com/NVIDIA/nvidia-docker)\n- TensorFlow2 22.11-py3+ NGC container\n- Supported GPUs:\n - [NVIDIA Volta architecture](https://www.nvidia.com/en-us/data-center/volta-gpu-architecture/)\n - [NVIDIA Turing architecture](https://www.nvidia.com/en-us/geforce/turing/)\n - [NVIDIA Ampere architecture](https://www.nvidia.com/en-us/data-center/nvidia-ampere-gpu-architecture/)\n\nFor more information about how to get started with NGC containers, see the following sections from the NVIDIA GPU Cloud Documentation and the Deep Learning Documentation:\n- [Getting Started Using NVIDIA GPU Cloud](https://docs.nvidia.com/ngc/ngc-getting-started-guide/index.html)\n- [Accessing And Pulling From The NGC Container Registry](https://docs.nvidia.com/deeplearning/frameworks/user-guide/index.html#accessing_registry)\n- Running [TensorFlow](https://docs.nvidia.com/deeplearning/frameworks/tensorflow-release-notes/running.html#running)\n \nFor those unable to use the PyTorch NGC container, to set up the required environment or create your own container, see the versioned [NVIDIA Container Support Matrix](https://docs.nvidia.com/deeplearning/frameworks/support-matrix/index.html).","hasSetup","quickStartGuide","To train your model using mixed or TF32 precision with Tensor Cores or using FP32, perform the following steps using the default parameters of the nnUNet model on the [Medical Segmentation Decathlon](http://medicaldecathlon.com/) dataset. For the specifics on training and inference, see the [Advanced](#advanced) section.\n\n1. Clone the repository.\n\nExecuting this command will create your local repository with all the code to run nnU-Net.\n```\ngit clone https://github.com/NVIDIA/DeepLearningExamples\ncd DeepLearningExamples/TensorFlow2/Segmentation/nnUNet\n```\n \n2. Build the nnU-Net TensorFlow2 NGC container.\n \nThis command will use the Dockerfile to create a Docker image named `nnunet`, downloading all the required components.\n\n```\ndocker build -t nnunet .\n```\n \nThe NGC container contains all the components optimized for usage on NVIDIA hardware.\n \n3. Start an interactive session in the NGC container to run preprocessing/training/inference.\n \nThe following command will launch the container and mount the `./data` directory as a volume to the `/data` directory inside the container, and `./results` directory to the `/results` directory in the container.\n \n```\nmkdir data results\ndocker run -it --privileged --runtime=nvidia --shm-size=8g --ulimit memlock=-1 --ulimit stack=67108864 --rm -v ${PWD}/data:/data -v ${PWD}/results:/results nnunet:latest /bin/bash\n```\n\n4. Prepare the BraTS (MSD 01 task) dataset.\n\nTo download and preprocess the data run:\n```\npython download.py --task 01\npython preprocess.py --task 01 --dim 3\npython preprocess.py --task 01 --dim 2\n```\n\nThen `ls /data` should print:\n```\n01_3d_tf2 01_2d_tf2 Task01_BrainTumour\n```\n\nFor the specifics on data preprocessing, see the [Getting the data](#getting-the-data) section.\n \n5. Start training.\n \nTraining can be started with:\n```\npython scripts/train.py --gpus --fold --dim [--amp]\n```\n\nTo see descriptions of the train script arguments run `python scripts/train.py --help`. You can customize the training process. For details, see the [Training process](#training-process) section.\n\n6. Start benchmarking.\n\nThe training and inference performance can be evaluated by using benchmarking scripts, such as:\n \n```\npython scripts/benchmark.py --mode {train,predict} --gpus --dim {2,3} --batch-size [--amp]\n```\n\nTo see descriptions of the benchmark script arguments run `python scripts/benchmark.py --help`.\n\n7. Start inference/predictions.\n \nInference can be started with:\n```\npython scripts/inference.py --data --dim --fold --ckpt-dir [--amp] [--tta] [--save-preds]\n```\n\nNote: You have to prepare either validation or test dataset to run this script by running `python preprocess.py --task 01 --dim {2,3} --exec_mode {val,test}`. After preprocessing inside a given task directory (e.g. `/data/01_3d` for task 01 and dim 3) it will create a `val` or `test` directory with preprocessed data ready for inference. Possible workflow:\n\n```\npython preprocess.py --task 01 --dim 3 --exec_mode val\npython scripts/inference.py --data /data/01_3d/val --dim 3 --fold 0 --ckpt-dir --amp --tta --save-preds\n```\n\nThen if you have labels for predicted images you can evaluate them with `evaluate.py` script. For example:\n\n```\npython evaluate.py --preds /results/preds_task_01_dim_3_fold_0_amp_tta --lbls /data/Task01_BrainTumour/labelsTr\n```\n\nTo see descriptions of the inference script arguments run `python scripts/inference.py --help`. You can customize the inference process. For details, see the [Inference process](#inference-process) section.\n\nNow that you have your model trained and evaluated, you can choose to compare your training results with our [Training accuracy results](#training-accuracy-results). You can also choose to benchmark yours performance to [Training performance benchmark](#training-performance-results), or [Inference performance benchmark](#inference-performance-results). Following the steps in these sections will ensure that you achieve the same accuracy and performance results as stated in the [Results](#results) section.","hasQuickStartGuide","advanced","The following sections provide greater details of the dataset, running training and inference, and the training results.\n\n### Scripts and sample code\n\nIn the root directory, the most important files are:\n\n* `main.py`: Entry point to the application. Runs training, evaluation, inference or benchmarking.\n* `preprocess.py`: Entry point to data preprocessing.\n* `download.py`: Downloads given dataset from [Medical Segmentation Decathlon](http://medicaldecathlon.com/).\n* `Dockerfile`: Container with the basic set of dependencies to run nnU-Net.\n* `requirements.txt:` Set of extra requirements for running nnU-Net.\n* `evaluate.py`: Compare predictions with ground truth and get the final score.\n \nThe `data_preprocessing` folder contains information about the data preprocessing used by nnU-Net. Its contents are:\n \n* `configs.py`: Defines dataset configuration like patch size or spacing.\n* `preprocessor.py`: Implements data preprocessing pipeline.\n\nThe `data_loading` folder contains information about the data-loading pipeline used by nnU-Net. Its contents are:\n \n* `data_module.py`: Defines a data module managing datasets and splits (similar to PyTorch Lightning `DataModule`)\n* `dali_loader.py`: Implements DALI data loading pipelines.\n* `utils.py`: Defines auxiliary functions used for data loading.\n\nThe `models` folder contains information about the building blocks of nnU-Net and the way they are assembled. Its contents are:\n \n* `layers.py`: Implements convolution blocks used by the U-Net template.\n* `nn_unet.py`: Implements training/validation/test logic and dynamic creation of U-Net architecture used by nnU-Net.\n* `sliding_window.py`: Implements sliding window inference used by evaluation and prediction loops.\n* `unet.py`: Implements the U-Net template.\n\nThe `runtime` folder contains information about training, inference, and evaluation logic. Its contents are:\n \n* `args.py`: Defines command line arguments.\n* `checkpoint.py`: Implements checkpoint saving.\n* `logging.py`: Defines logging utilities along with wandb.io integration.\n* `losses.py`: Implements loss functions.\n* `metrics.py`: Implements dice metric and metric management.\n* `run.py`: Implements training loop and inference and evaluation logic.\n* `utils.py`: Defines auxiliary functions used during runtime.\n\nOther folders included in the root directory are:\n\n* `images/`: Contains a model diagram.\n* `scripts/`: Provides scripts for training, benchmarking, and inference of nnU-Net.\n\n### Command-line options\n\nTo see the full list of available options and their descriptions, use the `-h` or `--help` command-line option, for example:\n\n`python main.py --help`\n\nThe following example output is printed when running the model:\n\n```\nusage: main.py [-h] [--exec-mode {train,evaluate,predict,export,nav}] [--gpus GPUS] [--data DATA] [--task TASK] [--dim {2,3}]\n [--seed SEED] [--benchmark] [--tta [BOOLEAN]] [--save-preds [BOOLEAN]] [--sw-benchmark [BOOLEAN]]\n [--results RESULTS] [--logname LOGNAME] [--quiet] [--use-dllogger [BOOLEAN]] [--amp [BOOLEAN]] [--xla [BOOLEAN]]\n [--read-roi [BOOLEAN]] [--batch-size BATCH_SIZE] [--learning-rate LEARNING_RATE] [--momentum MOMENTUM]\n [--scheduler {none,poly,cosine,cosine_annealing}] [--end-learning-rate END_LEARNING_RATE]\n [--cosine-annealing-first-cycle-steps COSINE_ANNEALING_FIRST_CYCLE_STEPS]\n [--cosine-annealing-peak-decay COSINE_ANNEALING_PEAK_DECAY] [--optimizer {sgd,adam,radam}]\n [--deep-supervision [BOOLEAN]] [--lookahead [BOOLEAN]] [--weight-decay WEIGHT_DECAY]\n [--loss-batch-reduction [BOOLEAN]] [--loss-include-background [BOOLEAN]] [--negative-slope NEGATIVE_SLOPE]\n [--norm {instance,batch,group,none}] [--ckpt-strategy {last_and_best,last_only,none}] [--ckpt-dir CKPT_DIR]\n [--saved-model-dir SAVED_MODEL_DIR] [--resume-training] [--load_sm [BOOLEAN]] [--validate [BOOLEAN]] [--nvol NVOL]\n [--oversampling OVERSAMPLING] [--num-workers NUM_WORKERS] [--sw-batch-size SW_BATCH_SIZE] [--overlap OVERLAP]\n [--blend {gaussian,constant}] [--nfolds NFOLDS] [--fold FOLD] [--epochs EPOCHS] [--skip-eval SKIP_EVAL]\n [--steps-per-epoch STEPS_PER_EPOCH] [--bench-steps BENCH_STEPS] [--warmup-steps WARMUP_STEPS]\n\noptional arguments:\n -h, --help show this help message and exit\n --exec-mode {train,evaluate,predict,export,nav}, --exec_mode {train,evaluate,predict,export,nav}\n Execution mode to run the model (default: train)\n --gpus GPUS\n --data DATA Path to data directory (default: /data)\n --task TASK Task number, MSD uses numbers 01-10 (default: 01)\n --dim {2,3} UNet dimension (default: 3)\n --seed SEED Random seed (default: None)\n --benchmark Run model benchmarking (default: False)\n --tta [BOOLEAN] Enable test time augmentation (default: False)\n --save-preds [BOOLEAN], --save_preds [BOOLEAN]\n Save predictions (default: False)\n --sw-benchmark [BOOLEAN], --sw_benchmark [BOOLEAN]\n --results RESULTS Path to results directory (default: /results)\n --logname LOGNAME DLLogger output filename (default: dllogger.json)\n --quiet Minimalize stdout/stderr output (default: False)\n --use-dllogger [BOOLEAN], --use_dllogger [BOOLEAN]\n Use DLLogger logging (default: True)\n --amp [BOOLEAN] Enable automatic mixed precision (default: False)\n --xla [BOOLEAN] Enable XLA compiling (default: False)\n --read-roi [BOOLEAN], --read_roi [BOOLEAN]\n Use DALI direct ROI loading feature (default: False)\n --batch-size BATCH_SIZE, --batch_size BATCH_SIZE\n Batch size (default: 2)\n --learning-rate LEARNING_RATE, --learning_rate LEARNING_RATE\n Learning rate (default: 0.0003)\n --momentum MOMENTUM Momentum factor (SGD only) (default: 0.99)\n --scheduler {none,poly,cosine,cosine_annealing}\n Learning rate scheduler (default: none)\n --end-learning-rate END_LEARNING_RATE\n End learning rate for poly scheduler (default: 5e-05)\n --cosine-annealing-first-cycle-steps COSINE_ANNEALING_FIRST_CYCLE_STEPS\n Length of a cosine decay cycle in steps, only with 'cosine_annealing' scheduler (default: 512)\n --cosine-annealing-peak-decay COSINE_ANNEALING_PEAK_DECAY\n Multiplier reducing initial learning rate (default: 0.95)\n --optimizer {sgd,adam,radam}\n Optimizer (default: adam)\n --deep-supervision [BOOLEAN], --deep_supervision [BOOLEAN]\n Use deep supervision. (default: False)\n --lookahead [BOOLEAN]\n Use Lookahead with the optimizer (default: False)\n --weight-decay WEIGHT_DECAY, --weight_decay WEIGHT_DECAY\n Weight decay (L2 penalty) (default: 0.0001)\n --loss-batch-reduction [BOOLEAN]\n Reduce batch dimension first during loss calculation (default: True)\n --loss-include-background [BOOLEAN]\n Include background class to loss calculation (default: False)\n --negative-slope NEGATIVE_SLOPE\n Negative slope for LeakyReLU (default: 0.01)\n --norm {instance,batch,group,none}\n Type of normalization layers (default: instance)\n --ckpt-strategy {last_and_best,last_only,none}\n Strategy how to save checkpoints (default: last_and_best)\n --ckpt-dir CKPT_DIR Path to checkpoint directory (default: /results/ckpt)\n --saved-model-dir SAVED_MODEL_DIR\n Path to saved model directory (for evaluation and prediction) (default: None)\n --resume-training, --resume_training\n Resume training from the last checkpoint (default: False)\n --load_sm [BOOLEAN] Load exported savedmodel (default: False)\n --validate [BOOLEAN] Validate exported savedmodel (default: False)\n --nvol NVOL Number of volumes which come into single batch size for 2D model (default: 2)\n --oversampling OVERSAMPLING\n Probability of crop to have some region with positive label (default: 0.33)\n --num-workers NUM_WORKERS\n Number of subprocesses to use for data loading (default: 8)\n --sw-batch-size SW_BATCH_SIZE\n Sliding window inference batch size (default: 2)\n --overlap OVERLAP Amount of overlap between scans during sliding window inference (default: 0.5)\n --blend {gaussian,constant}, --blend-mode {gaussian,constant}\n How to blend output of overlapping windows (default: gaussian)\n --nfolds NFOLDS Number of cross-validation folds (default: 5)\n --fold FOLD Fold number (default: 0)\n --epochs EPOCHS Number of epochs (default: 1000)\n --skip-eval SKIP_EVAL\n Skip evaluation for the first N epochs. (default: 0)\n --steps-per-epoch STEPS_PER_EPOCH\n Steps per epoch. By default ceil(training_dataset_size / batch_size / gpus) (default: None)\n --bench-steps BENCH_STEPS\n Number of benchmarked steps in total (default: 100)\n --warmup-steps WARMUP_STEPS\n Number of warmup steps before collecting benchmarking statistics (default: 25)\n```\n\n### Getting the data\n\nThe nnU-Net model was trained on the [Medical Segmentation Decathlon](http://medicaldecathlon.com/) datasets. All datasets are in Neuroimaging Informatics Technology Initiative (NIfTI) format.\n\n#### Dataset guidelines\n\nTo train nnU-Net you will need to preprocess your dataset as the first step with `preprocess.py` script. Run `python scripts/preprocess.py --help` to see descriptions of the preprocess script arguments.\n\nFor example to preprocess data for 3D U-Net run: `python preprocess.py --task 01 --dim 3`.\n\nIn `data_preprocessing/configs.py` for each [Medical Segmentation Decathlon](http://medicaldecathlon.com/) task, there are defined: patch sizes, precomputed spacings and statistics for CT datasets.\n\nThe preprocessing pipeline consists of the following steps:\n\n1. Cropping to the region of non-zero values.\n2. Resampling to the median voxel spacing of their respective dataset (exception for anisotropic datasets where the lowest resolution axis is selected to be the 10th percentile of the spacings).\n3. Padding volumes so that dimensions are at least as patch size.\n4. Normalizing:\n * For CT modalities the voxel values are clipped to 0.5 and 99.5 percentiles of the foreground voxels and then data is normalized with mean and standard deviation collected from foreground voxels.\n * For MRI modalities z-score normalization is applied.\n\n#### Multi-dataset\n\nIt is possible to run nnUNet on a custom dataset. If your dataset corresponds to [Medical Segmentation Decathlon](http://medicaldecathlon.com/) (i.e. data should be in `NIfTi` format and there should be `dataset.json` file where you need to provide fields: modality, labels, and at least one of training, test) you need to perform the following:\n\n1. Mount your dataset to `/data` directory.\n \n2. In `data_preprocessing/config.py`:\n - Add to the `task_dir` dictionary your dataset directory name. For example, for the Brain Tumour dataset, it corresponds to `\"01\": \"Task01_BrainTumour\"`.\n - Add the patch size that you want to use for training to the `patch_size` dictionary. For example, for the Brain Tumour dataset it corresponds to `\"01_3d\": [128, 128, 128]` for 3D U-Net and `\"01_2d\": [192, 160]` for 2D U-Net. There are three types of suffixes `_3d, _2d` corresponding to 3D UNet and 2D U-Net.\n\n3. Preprocess your data with `preprocess.py` scripts. For example, to preprocess the Brain Tumour dataset for 2D U-Net you should run `python preprocess.py --task 01 --dim 2`.\n\n### Training process\n\nThe model trains for at most `--epochs` epochs. After each epoch evaluation, the validation set is done and validation metrics are monitored for checkpoint updating\n(see `--ckpt-strategy` flag). Default training settings are:\n* The Adam optimizer with a learning rate of 0.0003 and weight decay of 0.0001.\n* Training batch size is set to 2.\n\nThis default parametrization is applied when running scripts from the `scripts/` directory and when running `main.py` without explicitly overriding these parameters. By default, using scripts from the `scripts` directory will not use AMP. To enable AMP, pass the `--amp` flag. AMP can be enabled for every mode of execution. However, a custom invocation of the `main.py` script will turn on AMP by default. To turn it off use `main.py --amp false`.\n\nThe default configuration minimizes a function `L = (1 - dice_coefficient) + cross_entropy` during training and reports achieved convergence as [dice coefficient](https://en.wikipedia.org/wiki/S%C3%B8rensen%E2%80%93Dice_coefficient) per class. The training, with a combination of dice and cross-entropy has been proven to achieve better convergence than training using only dice.\n\nThe training can be run without using the predefined scripts. The name of the training script is `main.py`. For example:\n\n```\npython main.py --exec-mode train --task 01 --fold 0\n```\n \nTraining artifacts will be saved to `/results` in the container. Some important artifacts are:\n* `/results/dllogger.json`: Collected dice scores and loss values evaluated after each epoch during training on a validation set.\n* `/results/ckpt/`: Saved checkpoints. By default, two checkpoints are saved - one after each epoch and one with the highest validation dice (saved in the `/results/ckpt/best/` subdirectory). You can change this behavior by modifying the `--ckpt-strategy` parameter.\n\nTo load the pretrained model, provide `--ckpt-dir ` and use `--resume-training` if you intend to continue training.\n\nTo use multi-GPU training with the `main.py` script prepend the command with `horovodrun -np `, for example with 8 GPUs use:\n\n```\nhorovodrun -np 8 python main.py --exec-mode train --task 01 --fold 0\n```\n### Inference process\n\nInference can be launched by passing the `--exec-mode predict` flag. For example:\n\n```\npython main.py --exec-mode predict --xla --task 01 --fold 0 --gpus 1 --amp --tta --save-preds --ckpt-dir \n```\n\nThe script will then:\n\n* Load the checkpoint from the directory specified by the `` directory\n* Run inference on the preprocessed validation dataset corresponding to fold 0\n* If `--save-preds` is provided then resulting masks in the NumPy format will be saved in the `/results` directory","hasAdvanced","performance","### Benchmarking\n\nThe following section shows how to run benchmarks to measure the model performance in training and inference modes.\n\n#### Training performance benchmark\n\nTo benchmark training, run the `scripts/benchmark.py` script with `--mode train`:\n\n```\npython scripts/benchmark.py --xla --mode train --gpus --dim {2,3} --batch-size [--amp]\n```\n\nFor example, to benchmark 3D U-Net training using mixed-precision on 8 GPUs with batch size of 2, run:\n\n```\npython scripts/benchmark.py --xla --mode train --gpus 8 --dim 3 --batch-size 2 --amp\n```\n\nEach of these scripts will by default run a warm-up for 100 iterations and then start benchmarking for another 100 steps.\nYou can adjust these settings with `--warmup-steps` and `--bench-steps` parameters.\n\nAt the end of the script, a line reporting the training throughput and latency will be printed.\n\n#### Inference performance benchmark\n\nTo benchmark inference, run the `scripts/benchmark.py` script with `--mode predict`:\n\n```\npython scripts/benchmark.py --xla --mode predict --gpus --dim {2,3} --batch-size [--amp]\n```\n\nFor example, to benchmark inference using mixed-precision for 3D U-Net on 1 GPU, with a batch size of 4, run:\n\n```\npython scripts/benchmark.py --xla --mode predict --gpus 1 --dim 3 --batch-size 4 --amp \n```\n\nEach of these scripts will by default run a warm-up for 100 iterations and then start benchmarking for another 100 steps.\nYou can adjust these settings with `--warmup-steps` and `--bench-steps` parameters.\n\nAt the end of the script, a line reporting the inference throughput and latency will be printed.\n\n*Note that this benchmark reports performance numbers for iterations over samples with fixed patch sizes.\nThe real inference process uses [sliding window](#glossary) for input images with arbitrary resolution and performance may vary for images with different resolutions.*\n\n### Results\n\nThe following sections provide details on how to achieve the same performance and accuracy in training and inference.\n\n#### Training accuracy results\n\n##### Training accuracy: NVIDIA DGX A100 (8xA100 80G)\n\nOur results were obtained by running the `python scripts/train.py --xla --gpus {1,8} --fold {0,1,2,3,4} --dim {2,3} --learning_rate lr [--amp] --seed n` training scripts and averaging results in the TensorFlow 22.11 NGC container on NVIDIA DGX with (8x A100 80G) GPUs.\n\n| Dimension | GPUs | Batch size / GPU | Dice - mixed precision | Accuracy - FP32 | Time to train - mixed precision | Time to train - TF32 | Time to train speedup (TF32 to mixed precision) \n|:-:|:-:|:--:|:-----:|:-----:|:--------:|:---------:|:----:|\n| 2 | 1 | 64 | 0.7312 | 0.7302 | 29 min | 40 min | 1.38 |\n| 2 | 8 | 64 | 0.7322 | 0.7310 | 8 min | 10 min | 1.22 |\n| 3 | 1 | 2 | 0.7435 | 0.7441 | 85 min | 153 min | 1.79 |\n| 3 | 8 | 2 | 0.7440 | 0.7438 | 19 min | 33 min | 1.69 |\n\nReported dice score is the average over 5 folds from the best run for grid search over learning rates {1e-4, 2e-4, ..., 9e-4} and seed {1, 3, 5}.\n\n##### Training accuracy: NVIDIA DGX-1 (8xV100 32G)\n\nOur results were obtained by running the `python scripts/train.py --xla --gpus {1,8} --fold {0,1,2,3,4} --dim {2,3} [--amp] --seed n` training scripts and averaging results in the TensorFlow 22.11 NGC container on NVIDIA DGX-1 with (8x V100 32G) GPUs.\n\n| Dimension | GPUs | Batch size / GPU | Dice - mixed precision | Accuracy - FP32 | Time to train - mixed precision | Time to train - FP32 | Time to train speedup (FP32 to mixed precision) \n|:-:|:-:|:--:|:-----:|:-----:|:---------:|:---------:|:----:|\n| 2 | 1 | 64 | 0.7315 | 0.7311 | 52 min | 102 min | 1.96 |\n| 2 | 8 | 64 | 0.7312 | 0.7316 | 12 min | 17 min | 1.41 |\n| 3 | 1 | 2 | 0.7435 | 0.7441 | 181 min | 580 min | 3.20 |\n| 3 | 8 | 2 | 0.7434 | 0.7440 | 35 min | 131 min | 3.74 |\n\nReported dice score is the average over 5 folds from the best run for grid search over learning rates {1e-4, 2e-4, ..., 9e-4} and seed {1, 3, 5}.\n\n#### Training performance results\n\n##### Training performance: NVIDIA DGX A100 (8xA100 80G)\n\nOur results were obtained by running the `python scripts/benchmark.py --xla --mode train --gpus {1,8} --dim {2,3} --batch-size [--amp]` training script in the NGC container on NVIDIA DGX A100 (8x A100 80G) GPUs. Performance numbers (in volumes per second) were averaged over an entire training epoch.\n\nNote: We recommend using `--bind` flag for multi-GPU settings to increase the throughput. To launch multi-GPU with `--bind` you will have to add `--horovod` e.g., `python scripts/benchmark.py --xla --mode train --gpus 8 --dim 3 --amp --batch-size 2 --bind --horovod` for the interactive session, or use regular command when launching with SLURM's sbatch.\n\n| Dimension | GPUs | Batch size / GPU | Throughput - mixed precision [img/s] | Throughput - TF32 [img/s] | Throughput speedup (TF32 - mixed precision) | Weak scaling - mixed precision | Weak scaling - TF32 |\n|:-:|:-:|:--:|:------:|:------:|:-----:|:-----:|:-----:|\n| 2 | 1 | 32 | 1347.19 | 748.56 | 1.80 | - | - |\n| 2 | 1 | 64 | 1662.8 | 804.23 | 2.07 | - | - |\n| 2 | 1 | 128 | 1844.7 | 881.87 | 2.09 | - | - |\n| 2 | 8 | 32 | 9056.45 | 5420.51 | 1.67 | 6.72 | 6.91 | \n| 2 | 8 | 64 | 11687.11 | 6250.52 | 1.87 | 7.03 | 7.49 |\n| 2 | 8 | 128 | 13679.76 | 6841.78 | 2.00 | 7.42 | 7.66 |\n| 3 | 1 | 1 | 27.02 | 11.63 | 2.32 | - | - |\n| 3 | 1 | 2 | 29.3 | 11.81 | 2.48 | - | - |\n| 3 | 1 | 4 | 31.87 | 12.17 | 2.62 | - | - |\n| 3 | 8 | 1 | 186.84 | 91.11 | 2.05 | 7.24 | 7.83 | \n| 3 | 8 | 2 | 219.34 | 92.91 | 2.36 | 7.77 | 7.87 | \n| 3 | 8 | 4 | 244.01 | 96.52 | 2.53 | 7.76 | 7.93 | \n\nTo achieve these same results, follow the steps in the [Quick Start Guide](#quick-start-guide).\n\n##### Training performance: NVIDIA DGX-1 (8xV100 32G)\n\nOur results were obtained by running the `python scripts/benchmark.py --xla --mode train --gpus {1,8} --dim {2,3} --batch-size [--amp]` training script in the TensorFlow 22.11 NGC container on NVIDIA DGX-1 with (8x V100 32G) GPUs. Performance numbers (in volumes per second) were averaged over an entire training epoch.\n\nNote: We recommend using `--bind` flag for multi-GPU settings to increase the throughput. To launch multi-GPU with `--bind` you will have to add `--horovod` e.g., `python scripts/benchmark.py --xla --mode train --gpus 8 --dim 3 --amp --batch-size 2 --bind --horovod` for the interactive session, or use regular command when launching with SLURM's sbatch.\n\n| Dimension | GPUs | Batch size / GPU | Throughput - mixed precision [img/s] | Throughput - FP32 [img/s] | Throughput speedup (FP32 - mixed precision) | Weak scaling - mixed precision | Weak scaling - FP32 |\n|:-:|:-:|:---:|:---------:|:-----------:|:--------:|:---------:|:-------------:|\n|\t2\t|\t1\t|\t32\t|\t697.36\t|\t312.51\t|\t2.23\t|\t-\t|\t-\t|\n|\t2\t|\t1\t|\t64\t|\t819.15\t|\t337.42\t|\t2.43\t|\t-\t|\t-\t|\n|\t2\t|\t1\t|\t128\t|\t894.94\t|\t352.32\t|\t2.54\t|\t-\t|\t-\t|\n|\t2\t|\t8\t|\t32\t|\t4355.65\t|\t2260.37\t|\t1.93\t|\t6.25\t|\t7.23\t|\n|\t2\t|\t8\t|\t64\t|\t5696.41\t|\t2585.65\t|\t2.20\t|\t6.95\t|\t7.66\t|\n|\t2\t|\t8\t|\t128\t|\t6714.96\t|\t2779.25\t|\t2.42\t|\t7.50\t|\t7.89\t|\n|\t3\t|\t1\t|\t1\t|\t12.15\t|\t2.08\t|\t5.84\t|\t-\t|\t-\t|\n|\t3\t|\t1\t|\t2\t|\t13.13\t|\t2.5\t|\t5.25\t|\t-\t|\t-\t|\n|\t3\t|\t8\t|\t1\t|\t82.62\t|\t16.59\t|\t4.98\t|\t6.80\t|\t7.98\t|\n|\t3\t|\t8\t|\t2\t|\t97.68\t|\t19.91\t|\t4.91\t|\t7.44\t|\t7.96\t|\n\nTo achieve these same results, follow the steps in the [Quick Start Guide](#quick-start-guide).\n\n#### Inference performance results\n\n##### Inference performance: NVIDIA DGX A100 (1xA100 80G)\n\nOur results were obtained by running the `python scripts/benchmark.py --xla --mode predict --dim {2,3} --batch-size [--amp]` inferencing benchmarking script in the TensorFlow 22.11 NGC container on NVIDIA DGX A100 (1x A100 80G) GPU.\n\nFP16\n\n| Dimension | Batch size |Resolution| Throughput Avg [img/s] | Latency Avg [ms] | Latency 90% [ms] | Latency 95% [ms] | Latency 99% [ms] |\n|:----------:|:---------:|:-------------:|:----------------------:|:----------------:|:----------------:|:----------------:|:----------------:|\n|\t2\t|\t32\t|\t192x160\t|\t1728.03\t|\t18.52\t| 22.55 | 23.18 | 24.82 |\n|\t2\t|\t64\t|\t192x160\t|\t4160.91\t|\t15.38\t| 17.49 | 18.53 | 19.88 |\n|\t2\t|\t128\t|\t192x160\t|\t4672.52\t|\t27.39\t| 27.68 | 27.79 | 27.87 |\n|\t3\t|\t1\t|\t128x128x128\t|\t78.2\t|\t12.79\t| 14.29 | 14.87 | 15.25 |\n|\t3\t|\t2\t|\t128x128x128\t|\t63.76\t|\t31.37\t| 36.07 | 40.02 | 42.44 |\n|\t3\t|\t4\t|\t128x128x128\t|\t83.17\t|\t48.1\t| 50.96 | 52.08 | 52.56 |\n\nTF32\n\n| Dimension | Batch size |Resolution| Throughput Avg [img/s] | Latency Avg [ms] | Latency 90% [ms] | Latency 95% [ms] | Latency 99% [ms] |\n|:----------:|:---------:|:-------------:|:----------------------:|:----------------:|:----------------:|:----------------:|:----------------:|\n|\t2\t|\t32\t|\t192x160\t|\t2067.63\t|\t15.48\t| 17.97 | 19.12 | 19.77 |\n|\t2\t|\t64\t|\t192x160\t|\t2447\t|\t26.15\t| 26.43 | 26.48 | 26.62 |\n|\t2\t|\t128\t|\t192x160\t|\t2514.75\t|\t50.9\t| 51.15 | 51.23 | 51.28 |\n|\t3\t|\t1\t|\t128x128x128\t|\t38.85\t|\t25.74\t| 26.04 | 26.19 | 27.41 |\n|\t3\t|\t2\t|\t128x128x128\t|\t40.1\t|\t49.87\t| 50.31 | 50.44 | 50.57 |\n|\t3\t|\t4\t|\t128x128x128\t|\t41.69\t|\t95.95\t| 97.09 | 97.41 | 98.03 |\n\nThroughput is reported in images per second. Latency is reported in milliseconds per batch.\nTo achieve these same results, follow the steps in the [Quick Start Guide](#quick-start-guide).\n\n##### Inference performance: NVIDIA DGX-1 (1xV100 32G)\n\nOur results were obtained by running the `python scripts/benchmark.py --mode predict --dim {2,3} --batch-size [--amp]` inferencing benchmarking script in the TensorFlow 22.11 NGC container on NVIDIA DGX-1 with (1x V100 32G) GPU.\n\nFP16\n \n| Dimension | Batch size |Resolution| Throughput Avg [img/s] | Latency Avg [ms] | Latency 90% [ms] | Latency 95% [ms] | Latency 99% [ms] |\n|:----------:|:---------:|:-------------:|:----------------------:|:----------------:|:----------------:|:----------------:|:----------------:|\n|\t2\t|\t32\t|\t192x160\t|\t1166.83\t|\t27.42\t| 28.76 | 28.91 | 29.16 |\n|\t2\t|\t64\t|\t192x160\t|\t2263.21\t|\t28.28\t| 30.63 | 31.83 | 32.5 |\n|\t2\t|\t128\t|\t192x160\t|\t2387.06\t|\t53.62\t| 53.97 | 54.07 | 54.3 |\n|\t3\t|\t1\t|\t128x128x128\t|\t36.87\t|\t27.12\t| 27.32 | 27.37 | 27.42 |\n|\t3\t|\t2\t|\t128x128x128\t|\t37.65\t|\t53.12\t| 53.49 | 53.59 | 53.71 |\n|\t3\t|\t4\t|\t128x128x128\t|\t38.8\t|\t103.11\t| 104.16 | 104.3 | 104.75 |\n\nFP32\n\n| Dimension | Batch size |Resolution| Throughput Avg [img/s] | Latency Avg [ms] | Latency 90% [ms] | Latency 95% [ms] | Latency 99% [ms] |\n|:----------:|:---------:|:-------------:|:----------------------:|:----------------:|:----------------:|:----------------:|:----------------:|\n|\t2\t|\t32\t|\t192x160\t|\t990.61\t|\t32.3\t| 32.46 | 32.51 | 32.78 |\n|\t2\t|\t64\t|\t192x160\t|\t1034.22\t|\t61.88\t| 62.19 | 62.32 | 62.56 |\n|\t2\t|\t128\t|\t192x160\t|\t1084.21\t|\t118.06\t| 118.45 | 118.6 | 118.95 |\n|\t3\t|\t1\t|\t128x128x128\t|\t9.65\t|\t103.62\t| 104.46 | 104.52 | 104.63 |\n|\t3\t|\t2\t|\t128x128x128\t|\t9.96\t|\t200.75\t| 202.51 | 202.74 | 202.86 |\n|\t3\t|\t4\t|\t128x128x128\t|\t10.13\t|\t394.74\t| 396.74 | 397.0 | 397.82 |\n\nThroughput is reported in images per second. Latency is reported in milliseconds per batch.\nTo achieve these same results, follow the steps in the [Quick Start Guide](#quick-start-guide).\n\n### Known issues\n\nThere are no known issues in this release.","hasPerformance",[],{"_87":241,"_116":242,"_117":-5,"_125":35,"_124":35,"_126":240,"_222":92},{"_87":29,"_116":29,"_117":-5,"_125":35,"_124":35,"_126":240,"_222":92},{"_87":115,"_116":115,"_117":-5,"_125":35,"_124":35,"_126":239,"_222":92},[],[],"-","Latest (22.11.0)","nvidia/dle/nnunet_tf2","key","nvidia/dle/resources/nnunet_tf2","resourceId","nnunet_tf2","shortDescription","An optimized, robust and self-adapting framework for U-Net based medical image segmentation","This resource is using open-source code maintained in github (see the quick-start-guide section) and available for download from [NGC](https://catalog.ngc.nvidia.com/orgs/nvidia/teams/dle/resources/nnunet-tf2)\n\n\nThe nnU-Net (\"no-new-Net\") refers to a robust and self-adapting framework for U-Net based medical image segmentation. This repository contains a nnU-Net implementation as described in the paper: [nnU-Net: Self-adapting Framework for U-Net-Based Medical Image Segmentation](https://arxiv.org/abs/1809.10486). \n\nThe differences between this nnU-net and [the original model](https://github.com/MIC-DKFZ/nnUNet) are:\n- Dynamic selection of patch size is not supported, and it has to be set in `data_preprocessing/configs.py` file.\n- Cascaded U-Net is not supported.\n- The following data augmentations are not used: rotation, simulation of low resolution, gamma augmentation.\n\nThis model is trained with mixed precision using Tensor Cores on Volta, Turing, and the NVIDIA Ampere GPU architectures. Therefore, researchers can get results 2x faster than training without Tensor Cores, while experiencing the benefits of mixed precision training. This model is tested against each NGC monthly container release to ensure consistent accuracy and performance over time.\n\n### Model architecture\n\nThe nnU-Net allows training two types of networks: 2D U-Net and 3D U-Net to perform semantic segmentation of 2D or 3D images, with high accuracy and performance.\n\nThe following figure shows the architecture of the 3D U-Net model and its different components. U-Net is composed of a contractive and an expanding path, that aims at building a bottleneck in its centermost part through a combination of convolution, instance norm, and leaky ReLU operations. After this bottleneck, the image is reconstructed through a combination of convolutions and upsampling. Skip connections are added with the goal of helping the backward flow of gradients to improve the training.\n\n

\n \n*Figure 1: The 3D U-Net architecture*\n\n### Default configuration\n\nAll convolution blocks in U-Net in both encoder and decoder are using two convolution layers followed by instance normalization and a leaky ReLU nonlinearity. For downsampling, we are using stride convolution whereas transposed convolution is used for upsampling.\n\nAll models were trained with the Adam optimizer. For loss function we use the average of [cross-entropy](https://en.wikipedia.org/wiki/Cross_entropy) and [dice coefficient](https://en.wikipedia.org/wiki/S%C3%B8rensen%E2%80%93Dice_coefficient).\n\nUsed data augmentation: crop with oversampling the foreground class, mirroring, zoom, Gaussian noise, Gaussian blur, brightness.\n\n### Feature support matrix\n\nThe following features are supported by this model: \n\n| Feature | nnUNet \n|-----------------------|-------------------------- \n|[DALI](https://docs.nvidia.com/deeplearning/dali/release-notes/index.html) | Yes\n|Automatic mixed precision (AMP) | Yes \n|Horovod Multi-GPU (NCCL) | Yes \n|[XLA](https://www.tensorflow.org/xla) | Yes\n\n#### Features\n\n**DALI**\nNVIDIA Data Loading Library (DALI) is a collection of optimized building blocks, and an execution engine, to speed up the pre-processing of the input data for deep learning applications. DALI provides both the performance and the flexibility for accelerating different data pipelines as a single library. This single library can then be integrated into different deep learning training and inference applications. For details, refer to example sources in this repository or refer to the [DALI documentation](https://docs.nvidia.com/deeplearning/dali/index.html).\n\n**Automatic Mixed Precision (AMP)**\nComputation graphs can be modified by TensorFlow during runtime to support mixed precision training, which allows using FP16 training with FP32 master weights. A detailed explanation of mixed precision can be found in the next section.\n \n**Multi-GPU training with Horovod**\nHorovod is a distributed training framework for TensorFlow, Keras, PyTorch, and MXNet. The goal of Horovod is to make distributed deep learning fast and easy to use. For more information about how to get started with Horovod, refer to the [Horovod: Official repository](https://github.com/horovod/horovod).\nOur model uses Horovod to implement efficient multi-GPU training with NCCL. For details, refer to example scripts in this repository or refer to the [TensorFlow tutorial](https://github.com/horovod/horovod/#usage).\n\n**XLA**\nXLA (Accelerated Linear Algebra) is a compiler that can speed up TensorFlow networks by model-specific optimizations i.e. fusing many GPU operations together.\nOperations fused into a single GPU kernel do not have to use extra memory to store intermediate values by keeping them in GPU registers, thus reducing memory operations and improving performance. For details refer to the [TensorFlow documentation](https://www.tensorflow.org/xla).\n\n### Mixed precision training\n\nMixed precision is the combined use of different numerical precisions in a computational method. [Mixed precision](https://arxiv.org/abs/1710.03740) training offers significant computational speedup by performing operations in the half-precision format while storing minimal information in single-precision to keep as much information as possible in critical parts of the network. Since the introduction of [Tensor Cores](https://developer.nvidia.com/tensor-cores) in Volta, and following with both the Turing and Ampere architectures, significant training speedups are experienced by switching to mixed precision -- up to 3x speedup on the most intense model architectures. Using mixed precision training requires two steps:\n\n1. Porting the model to use the FP16 data type where appropriate.\n2. Adding loss scaling to preserve small gradient values.\n\nThis can now be achieved using Automatic Mixed Precision (AMP) for TensorFlow to enable the full [mixed precision methodology](https://docs.nvidia.com/deeplearning/sdk/mixed-precision-training/index.html#tensorflow) in your existing TensorFlow model code. AMP enables mixed precision training on NVIDIA Volta, NVIDIA Turing, and NVIDIA Ampere GPU architectures automatically. The TensorFlow framework code makes all necessary model changes internally.\n\nIn TF-AMP, the computational graph is optimized to use as few casts as necessary and maximize the use of FP16, and the loss scaling is automatically applied inside of supported optimizers. AMP can be configured to work with the existing tf.contrib loss scaling manager by disabling the AMP scaling with a single environment variable to perform only the automatic mixed-precision optimization. It accomplishes this by automatically rewriting all computation graphs with the necessary operations to enable mixed precision training and automatic loss scaling.\n\nFor information about:\n* How to train using mixed precision, see the [Mixed Precision Training](https://arxiv.org/abs/1710.03740) paper and [Training With Mixed Precision](https://docs.nvidia.com/deeplearning/performance/mixed-precision-training/index.html) documentation.\n* Techniques used for mixed precision training, see the [Mixed-Precision Training of Deep Neural Networks](https://devblogs.nvidia.com/mixed-precision-training-deep-neural-networks/) blog.\n\n#### Enabling mixed precision\n\nMixed precision is enabled in TensorFlow by using the Automatic Mixed Precision (TF-AMP) extension which casts variables to half-precision upon retrieval, while storing variables in single-precision format. Furthermore, to preserve small gradient magnitudes in backpropagation, a [loss scaling](https://docs.nvidia.com/deeplearning/sdk/mixed-precision-training/index.html#lossscaling) step must be included when applying gradients. In TensorFlow, loss scaling can be applied statically by using simple multiplication of loss by a constant value or automatically, by TF-AMP. Automatic mixed precision makes all the adjustments internally in TensorFlow, providing two benefits over manual operations. First, programmers need not modify network model code, reducing development and maintenance efforts. Second, using AMP maintains forward and backward compatibility with all the APIs for defining and running TensorFlow models.\n\nExample nnU-Net scripts for training, inference, and benchmarking from the `scripts/` directory enable mixed precision if the `--amp` command line flag is used.\n\nInternally, mixed precision is enabled by setting `keras.mixed_precision` policy to `mixed_float16`. Additionally, our custom training loop uses a `LossScaleOptimizer` wrapper for the optimizer. For more information see the [Mixed precision guide](#mixed-precision-training).\n\n#### TF32\n\nTensorFloat-32 (TF32) is the new math mode in [NVIDIA A100](https://www.nvidia.com/en-us/data-center/a100/) GPUs for handling the matrix math, also called tensor operations. TF32 running on Tensor Cores in A100 GPUs can provide up to 10x speedups compared to single-precision floating-point math (FP32) on NVIDIA Volta GPUs. \n\nTF32 Tensor Cores can speed up networks using FP32, typically with no loss of accuracy. It is more robust than FP16 for models which require a high dynamic range for weights or activations.\n\nFor more information, refer to the [TensorFloat-32 in the A100 GPU Accelerates AI Training, HPC up to 20x](https://blogs.nvidia.com/blog/2020/05/14/tensorfloat-32-precision-format/) blog post.\n\nTF32 is supported in the NVIDIA Ampere GPU architecture and is enabled by default.\n\n### Glossary\n\n**Deep supervision**\nDeep supervision is a technique that adds auxiliary loss outputs to the U-Net decoder layers. For nnU-Net, we add auxiliary losses to the three latest decoder levels. The final loss is a weighted average of the obtained loss values. Deep supervision can be enabled by adding the `--deep-supervision` flag.\n\n**Test time augmentation**\nTest time augmentation is an inference technique that averages predictions from augmented images with its prediction. As a result, predictions are more accurate, but with the cost of a slower inference process. For nnU-Net, we use all possible flip combinations for image augmenting. Test time augmentation can be enabled by adding the `--tta` flag to the training or inference script invocation.\n\n**Sliding window inference**\nDuring inference, this method replaces an arbitrary resolution input image with a batch of overlapping windows, that cover the whole input. After passing this batch through the network a prediction with the original resolution is reassembled. Predicted values inside overlapped regions are obtained from a weighted average.\nOverlap ratio and weights for the average (i.e. blending mode) can be adjusted with the `--overlap` and `--blend-mode` options.","displayName","nnU-Net for TensorFlow2",["D",1670257653890],"publisher","NVIDIA Deep Learning Examples","latestVersion","collectionCounts","imageUrl","/api/image/nvidia/dle/resources/nnunet_tf2","rawImageUrl","https://assets.nvidiagrid.net/ngc/logos/Deep-Learning-Examples.png","canGuestDownload","latestVersionSizeInBytes","licenseTerms","isMultinodeEnabled","isMultiArchitecture","bias","explainability","privacy","safetyAndSecurity","featureLabels",[],"displayLabels",[],"policyLabels",[],"systemLabels",[],"products",[],"hasNimMetadata",{"_158":1264,"_160":1265},"ngcApps",[1217],"csaApiBaseUrl","https://www.nvidia.com/ai-agents/csa-agent/dcsm","datadogRum",{"_1198":1199,"_1200":1201,"_1202":1203,"_1204":1205,"_1206":293,"_111":319,"_1207":1208,"_1209":60,"_1210":92,"_1211":92,"_1212":1213,"_1214":92,"_1215":1216},"amplitude",{"_1195":1196,"_832":-7,"_1197":35},"disableJs","environment","prod","featureFlagsBlob","JTdCJTIydGVzdEZsYWclMjIlM0FmYWxzZSUyQyUyMnNob3dTZWFyY2hMYWJlbFNldEZpbHRlcnMlMjIlM0FmYWxzZSUyQyUyMmVuYWJsZUNzYVdpZGdldCUyMiUzQWZhbHNlJTJDJTIyZW5hYmxlUHJvZHVjdEdyb3VwaW5nQnlTdWJzY3JpcHRpb24lMjIlM0FmYWxzZSUyQyUyMmNvbnRhaW5lclNjYW5uaW5nUmVkZXNpZ25NMiUyMiUzQWZhbHNlJTdE","featureFlagsGuarded","isDevLikeEnv","isPlaywright","isMobileDevice","categorizedOrgs",{"_1189":1190,"_1191":1192,"_1193":1194},"currentOrgContext","currentContextRoles",[],"enterpriseOrgs",[],"manageableExternalUserGroups",[],"ngcApiOrigin","ngcHelmOrigin","origin","https://catalog.ngc.nvidia.com","usesMocks","theme","year",2026,"now",1789162842977,"1.15.0","defaultSearchSuggestions",{"_1072":400,"_1073":1074,"_1075":1076,"_1077":1078},"announcements",[],"consentState",{"_357":1062,"_1063":92,"_1064":35,"_1065":35,"_1066":1067},"consentConfig",{"_339":340,"_341":342,"_343":344,"_345":346,"_347":348,"_349":350,"_351":352,"_353":354,"_355":356,"_357":358},"showContactAdmin",1789162842978,[],[],[333,334,335,336,337,338],{"_49":64,"_51":75,"_53":153,"_55":-5,"_57":-5,"_59":60},{"_49":64,"_51":71,"_53":153,"_55":-5,"_57":-5,"_59":60},{"_49":64,"_51":69,"_53":153,"_55":-5,"_57":-5,"_59":60},{"_49":64,"_51":65,"_53":153,"_55":-5,"_57":-5,"_59":60},{"_49":61,"_51":62,"_53":153,"_55":-5,"_57":-5,"_59":60},{"_49":50,"_51":52,"_53":153,"_55":-5,"_57":-5,"_59":60},"mainInfoText","We and our third-party partners (including social media, advertising, and analytics partners) use cookies and other tracking technologies to collect, store, monitor, and process certain information about you when you visit our website. The information collected might relate to you, your preferences, or your device. We use that information to make the site work, analyze performance and traffic on our website, provide a more personalized web experience, and assist in our marketing efforts.

Under certain privacy laws, you have the right to direct us not to \"sell\" or \"share\" your personal information for targeted advertising. To opt-out of the \"sale\" and \"sharing\" of personal information through cookies, you must opt-out of optional cookies using the toggles below. To opt out of the \"sale\" and \"sharing\" of data collected by other means (e.g., online forms) you must also update your data sharing preferences through the NVIDIA Preference Center.

Click on the different category headings below to find out more and change the settings according to your preference. You cannot opt out of Required Cookies as they are deployed to ensure the proper functioning of our website (such as prompting the cookie banner and remembering your settings, etc.). By clicking \"Save and Accept\" or \"Decline All\" at the bottom, you consent to the use of cookies and other tools as described in our Cookie Policy in accordance with your settings and accept our Terms of Service (which contains important waivers). For more information about our privacy practices, please see our Privacy Policy.","mainText","Cookie Settings","alertNoticeText",{"_1054":1055,"_1056":1057,"_1058":1059,"_1060":1061},"bannerAcceptLabel","Accept All","bannerRejectLabel","Reject Optional","bannerSettingsLabel","Manage Settings","preferenceSaveLabel","Save Choices","preferenceRejectLabel","Decline All","alwaysActiveText","Always Active","groups",[359,360,361,362,363],{"_364":744,"_116":745,"_216":746,"_368":92,"_369":18,"_371":747,"_373":748},{"_364":647,"_116":648,"_216":649,"_368":35,"_369":63,"_371":650,"_373":651},{"_364":555,"_116":556,"_216":557,"_368":35,"_369":58,"_371":558,"_373":559},{"_364":375,"_116":376,"_216":377,"_368":35,"_369":378,"_371":379,"_373":380},{"_364":365,"_116":366,"_216":367,"_368":35,"_369":370,"_371":372,"_373":374},"optanonGroupId","C0005","Social Media Cookies","These cookies are set by a range of social media services that we have added to the site to enable you to share our content with your friends and networks. They are capable of tracking your browser across other sites and building up a profile of your interests. This may impact the content and messages you see on other websites you visit. If you do not allow these cookies you may not be able to use or see these sharing tools.","alwaysActive","order",11,"cookies",[],"hosts",[],"C0004","Advertising Cookies","These cookies record your visit to our websites, the pages you have visited and the links you have followed to influence the advertisements that you see on other websites. These cookies and the information they collect may be managed by other companies, including our advertising partners, and may be used to build a profile of your interests and show you relevant advertising on other sites. We and our advertising partners will use this information to make our websites and the advertising displayed on it, more relevant to your interests.",5,[497,498,499,500,501,502,503,504,505,506,507,508,509,510,511,512,513,514,515,516,517,518],[381,382,383,384,385,386,387,388,389,390,391,392,393,394,395,396,397],{"_398":492,"_251":492,"_216":400,"_371":493},{"_398":488,"_251":488,"_216":400,"_371":489},{"_398":483,"_251":483,"_216":400,"_371":484},{"_398":479,"_251":479,"_216":400,"_371":480},{"_398":474,"_251":474,"_216":400,"_371":475},{"_398":469,"_251":469,"_216":400,"_371":470},{"_398":461,"_251":461,"_216":400,"_371":462},{"_398":453,"_251":453,"_216":400,"_371":454},{"_398":448,"_251":448,"_216":400,"_371":449},{"_398":443,"_251":443,"_216":400,"_371":444},{"_398":438,"_251":438,"_216":400,"_371":439},{"_398":433,"_251":433,"_216":400,"_371":434},{"_398":428,"_251":428,"_216":400,"_371":429},{"_398":421,"_251":421,"_216":400,"_371":422},{"_398":416,"_251":416,"_216":400,"_371":417},{"_398":409,"_251":409,"_216":400,"_371":410},{"_398":399,"_251":399,"_216":400,"_371":401},"hostName","amazon-adsystem.com","",[402,403],{"_116":408,"_405":399,"_406":35,"_216":407},{"_116":404,"_405":399,"_406":35,"_216":407},"ad-privacy","host","isSession","This domain is owned by online retailer Amazon and is used as part of its affiliate marketing programme.","ad-id","doubleclick.net",[411,412],{"_116":415,"_405":409,"_406":35,"_216":414},{"_116":413,"_405":409,"_406":35,"_216":414},"IDE","This domain is owned by Doubleclick (Google). The main business activity is: Doubleclick is Googles real time bidding advertising exchange","test_cookie","bing.com",[418],{"_116":419,"_405":416,"_406":35,"_216":420},"MUID","This domain is owned by Mircosoft - it is the site for the search engine Bing.","assets.adobedtm.com",[423,424],{"_116":427,"_405":421,"_406":35,"_216":426},{"_116":425,"_405":421,"_406":35,"_216":426},"kndctr_F207D74D549850760A4C98C6_AdobeOrg_*","If you know anything about this host and its uses please add your contributions here. ","kndctr_F207D74D549850760A4C98C6_AdobeOrg_consent","tiktok.com",[430],{"_116":431,"_405":428,"_406":35,"_216":432},"_ttp","TikTok is a social media platform owned by ByteDance, known for short-form video sharing. The service is used to create and share videos, and it utilizes cookies for various purposes including advertising and analytics.","t.co",[435],{"_116":436,"_405":433,"_406":35,"_216":437},"muc_ads","This domain is owned by Twitter. It is used as a URL shortening service, which allows Twitter to collect data on how links are shared.","snowplow.com",[440],{"_116":441,"_405":438,"_406":35,"_216":442},"sp","Testing","google.com",[445],{"_116":446,"_405":443,"_406":35,"_216":447},"NID","This domain is owned by Google Inc. Although Google is primarily known as a search engine, the company provides a diverse range of products and services. Its main source of revenue however is advertising. Google tracks users extensively both through its own products and sites, and the numerous technologies embedded into many millions of websites around the world. It uses the data gathered from most of these services to profile the interests of web users and sell advertising space to organisations based on such interest profiles as well as aligning adverts to the content on the pages where its customer's adverts appear.","linkedin.com",[450],{"_116":451,"_405":448,"_406":35,"_216":452},"bcookie","This domain is owned by LinkedIn, the business networking platform. It typically acts as a third party host where website owners have placed one of its content sharing buttons in their pages, although its content and services can be embedded in other ways. Although such buttons add functionality to the website they are on, cookies are set regardless of whether or not the visitor has an active Linkedin profile, or agreed to their terms and conditions. For this reason it is classified as a primarily tracking/targeting domain.","twitter.com",[455,456],{"_116":459,"_405":453,"_406":35,"_216":460},{"_116":457,"_405":453,"_406":35,"_216":458},"personalization_id","This domain is owned by Twitter. The main business activity is: Social Networking Services. Where twitter acts as a third party host, it collects data through a range of plug-ins and integrations, that is primarily used for tracking and targeting.","guest_*","This domain is owned by X (formerly Twitter). The main business activity is: Social Networking Services. Where X (formerly Twitter) acts as a third party host, it collects data through a range of plug-ins and integrations, that is primarily used for tracking and targeting.","youtube.com",[463,464],{"_116":467,"_405":461,"_406":35,"_216":468},{"_116":465,"_405":461,"_406":35,"_216":466},"VISITOR_INFO1_LIVE","This cookie is used as a unique identifier to track viewing of videos","__Secure-*","YouTube is a Google owned platform for hosting and sharing videos. YouTube collects user data through videos embedded in websites, which is aggregated with profile data from other Google services in order to display targeted advertising to web visitors across a broad range of their own and other websites.","demdex.net",[471],{"_116":472,"_405":469,"_406":35,"_216":473},"demdex","This cookie helps Adobe Audience Manger perform basic functions such as visitor identification, ID synchronization, segmentation, modeling, reporting, etc.","im-apps.net",[476],{"_116":477,"_405":474,"_406":35,"_216":478},"receive-cookie-deprecation","This domain is owned by im-apps.net which provides information and resources about various instant messaging (IM) applications to help users make informed decisions based on their communication needs.","influ2.com",[481],{"_116":482,"_405":479,"_406":35,"_216":442},"R","6sc.co",[485],{"_116":486,"_405":483,"_406":35,"_216":487},"6suuid","This domain is associated with the company 6Sense Insights, Inc., which provides predictive analytics and account engagement services. It is primarily used for tracking and targeting in marketing automation and advertising.","hushly.com",[490],{"_116":491,"_405":488,"_406":35,"_216":442},"_hly_gvid","t.contentsquare.net",[494],{"_116":495,"_405":492,"_406":92,"_216":496},"_cs_*","This is a session cookie. This is a pattern type cookie where a random 13-digit number is appended to the prefix _cs_.",{"_116":553,"_405":520,"_406":35,"_216":554},{"_116":551,"_405":520,"_406":35,"_216":552},{"_116":549,"_405":520,"_406":35,"_216":550},{"_116":431,"_405":520,"_406":35,"_216":548},{"_116":525,"_405":520,"_406":92,"_216":527},{"_116":547,"_405":520,"_406":35,"_216":400},{"_116":546,"_405":523,"_406":35,"_216":400},{"_116":544,"_405":520,"_406":35,"_216":545},{"_116":542,"_405":520,"_406":35,"_216":543},{"_116":541,"_405":520,"_406":35,"_216":400},{"_116":539,"_405":520,"_406":35,"_216":540},{"_116":525,"_405":538,"_406":92,"_216":527},{"_116":536,"_405":523,"_406":35,"_216":537},{"_116":536,"_405":526,"_406":35,"_216":537},{"_116":495,"_405":520,"_406":35,"_216":496},{"_116":534,"_405":520,"_406":35,"_216":535},{"_116":441,"_405":532,"_406":35,"_216":533},{"_116":530,"_405":520,"_406":35,"_216":531},{"_116":528,"_405":520,"_406":35,"_216":529},{"_116":525,"_405":526,"_406":92,"_216":527},{"_116":522,"_405":523,"_406":35,"_216":524},{"_116":519,"_405":520,"_406":92,"_216":521},"AMCVS_xxxxxAdobeOrg","nvidia.com","This is a pattern type cookie name associated with Adobe Marketing Cloud. It stores a unique visitor identifier, and uses an organisation identifier to allow a company to track users across their domains and services.This is a pattern type cookie name associated with Adobe Marketing Cloud. It stores a unique visitor identifier, and uses an organisation identifier.","_rdt_pn","www.nvidia.com","This cookie name is associated with www.mind.org.uk which is used for marketing purposes.","AMP_*","build.nvidia.com","Supports Amplitude product analytics.","_gcl_au","Used by Google AdSense for experimenting with advertisement efficiency across websites using their services","_mkto_trk","This cookie is associated with an email marketing service provided by Marketo. This tracking cookie allows a website to link visitor behaviour to the recipient of an email marketing campaign, to measure campaign effectiveness.","investor.nvidia.com","This cookie is associated with Eventbrite and is used to deliver content tailored to the end user's interests and improve content creation. This cookie is also used for event-booking purposes.","_gat_gtag_UA_140825383_2","Google Analytics","_rdt_uuid","This cookie is set by Reddit and is used for tracking the views of embedded videos and performance of the advertisements.","catalog.ngc.nvidia.com","_uetsid","This cookie is used by Bing to determine what ads should be shown that may be relevant to the end user perusing the site. ","kndctr_F207D74D549850760A4C98C6_AdobeOrg_bc_session_id","ttcsid_XXXXXXXXXXXXXXXXXXXXCategory","This cookie name is associated with TikTok which is used to track your activity on the platform, including how you interact with their embedded services on other websites.This cookie is associated with TikTok. It helps track your activity on the platform, including how you interact with their embedded services on other websites.","_fbp","Used by Facebook to deliver a series of advertisement products such as real time bidding from third party advertisers","_zitok","_twpid","This cookie is typically associated with the use of Taptap Digital, a company specializing in digital advertising. It is used for tracking and ad personalization purposes.","_tt_enable_cookie","This cookie name is associated with TikTok which is used to collect data about behavior and purchases on the website and to measure the effect of advertising. It is also used to track website activity to help optimize advertising efforts.","AMCV_xxxxxAdobeOrg","This cookie is used to identify a unique visitor for tracking purposes by Adobe Experience Cloud.This is a pattern type cookie name associated with Adobe Marketing Cloud. It stores a unique visitor identifier, and uses an organisation identifier to allow a company to track users across their domains and services.","ttcsid","This cookie name is associated with TikTok which is used to enable TikTok Ads session tracking by assigning a unique session ID to users who click on TikTok ads, allowing for the attribution of user behavior and conversions within the same session.This cookie is associated with TikTok. It helps track your activity on the platform, including how you interact with their embedded services on other websites.","C0003","Personalization Cookies","These cookies collect data about how you have interacted with our website to help us improve your web experience, such as which pages you have visited. These cookies may store a unique ID so that our system will remember you when you return. They may be set by us or by third party providers whose services we have added to our pages. These cookies enable us to provide enhanced website functionality and personalization as well as make the marketing messages we send to you more relevant to your interests. If you do not allow these cookies, then some or all of these services may not function properly.",[589,590,591,592,593,594,595,596,597,598,599,600,601,602,603,604,605,606,607,608,609,610,611,612,613],[560,561,562,563,564],{"_398":474,"_251":474,"_216":400,"_371":584},{"_398":448,"_251":448,"_216":400,"_371":579},{"_398":574,"_251":574,"_216":400,"_371":575},{"_398":569,"_251":569,"_216":400,"_371":570},{"_398":565,"_251":565,"_216":400,"_371":566},"nvidia.partners",[567],{"_116":568,"_405":565,"_406":35,"_216":442},"edge","app.hushly.com",[571],{"_116":572,"_405":569,"_406":35,"_216":573},"*CORS","This domain is associated with Hushly, a platform that provides lead conversion solutions for content marketing by using personalization and content engagement strategies.","github.com",[576],{"_116":577,"_405":574,"_406":35,"_216":578},"_octo","This domain is owned by Github. The main business activity is: Source code repository and collaboration platform",[580,581],{"_116":583,"_405":448,"_406":35,"_216":452},{"_116":582,"_405":448,"_406":35,"_216":452},"lidc","li_gc",[585,586],{"_116":588,"_405":474,"_406":35,"_216":478},{"_116":587,"_405":474,"_406":35,"_216":478},"imid_*","imid",{"_116":614,"_405":523,"_406":35,"_216":616},{"_116":645,"_405":520,"_406":35,"_216":646},{"_116":614,"_405":644,"_406":35,"_216":616},{"_116":642,"_405":520,"_406":92,"_216":643},{"_116":640,"_405":520,"_406":35,"_216":641},{"_116":639,"_405":622,"_406":92,"_216":400},{"_116":614,"_405":618,"_406":35,"_216":616},{"_116":637,"_405":520,"_406":35,"_216":638},{"_116":614,"_405":526,"_406":35,"_216":616},{"_116":625,"_405":523,"_406":35,"_216":400},{"_116":614,"_405":636,"_406":35,"_216":616},{"_116":634,"_405":520,"_406":35,"_216":635},{"_116":631,"_405":632,"_406":35,"_216":633},{"_116":629,"_405":630,"_406":35,"_216":400},{"_116":614,"_405":626,"_406":35,"_216":616},{"_116":628,"_405":520,"_406":35,"_216":400},{"_116":614,"_405":627,"_406":35,"_216":616},{"_116":625,"_405":626,"_406":35,"_216":400},{"_116":624,"_405":622,"_406":92,"_216":400},{"_116":614,"_405":623,"_406":35,"_216":616},{"_116":614,"_405":622,"_406":35,"_216":616},{"_116":614,"_405":621,"_406":35,"_216":616},{"_116":620,"_405":520,"_406":92,"_216":616},{"_116":617,"_405":618,"_406":92,"_216":619},{"_116":614,"_405":615,"_406":35,"_216":616},"_gd_*","info.nvidia.com","This cookie contains no identifying information and is used for diagnostic purposes by Impact Radius, a service we use to track signup commissions for our affiliate program. We have categorized it under Functionality because it is necessary for the operation of our affiliate program.","nv_country_code","developer.nvidia.com","This cookie is likely used to store the country code of the visitor to provide localized content and possibly for legal compliance with regional data privacy laws.","_gdxxxxxxxxxxxxx","nvidianews.nvidia.com","docs.api.nvidia.com","research.nvidia.com","readme_library","wcmmode","ai-podcast.nvidia.com","resources.nvidia.com","geo_locator","openedx-language-preference","learn.nvidia.com","doxygen_width","raytracing-docs.nvidia.com","This cookie is likely used to remember the width preferences of a sidebar or interface element within a documentation platform like Doxygen.","_sfid_eb78","This cookie is categorized as Functionality Cookie.","crowdsource.nvidia.com","mbox","Cookie set by Adobe SiteCatalyst analytics software - used for measuring the performance of page content using A/B split testing","readme_language","vid","This cookie name 'vid' is used for tracking the search request made by the user on the 'Find a practitioner' page.","at_check","Used by Adobe Target to check if cookies are enabled/supported on the browser","blogs.nvidia.com","nvHeadless","This cookie is likely used to track if a user is using the site in a 'headless' non-GUI mode, which could be for accessibility reasons or for automated processes.","C0002","Performance Cookies","These cookies are used to provide quantitative measures of our website visitors, such as the number of times you visit, time on page, your mouse movements, scrolling, clicks and keystroke activity on the websites; other browsing, search, or product research behavior; and what brought you to our site. These cookies may store a unique ID so that our system will remember you when you return. Information collected with these cookies is used to measure and find ways to improve website performance.",[672,673,674,675,676,677,678,679,680,681,682,683,684,685,686,687,688,689,690,691,692,693,694,695,696,697,698,699,700],[652,653,654,655],{"_398":667,"_251":667,"_216":400,"_371":668},{"_398":662,"_251":662,"_216":400,"_371":663},{"_398":461,"_251":461,"_216":400,"_371":659},{"_398":574,"_251":574,"_216":400,"_371":656},[657],{"_116":658,"_405":574,"_406":92,"_216":578},"_gh_sess",[660],{"_116":661,"_405":461,"_406":92,"_216":468},"YSC","bizible.com",[664],{"_116":665,"_405":662,"_406":35,"_216":666},"_BUID","Bizible is a marketing attribution software that tracks the performance of marketing efforts and connects them to revenue, primarily used for B2B environments.","a.usbrowserspeed.com",[669],{"_116":670,"_405":667,"_406":35,"_216":671},"tuid","This domain is owned by Ookla which provides internet speed testing services through its Speedtest application.",{"_116":742,"_405":520,"_406":35,"_216":743},{"_116":727,"_405":741,"_406":92,"_216":728},{"_116":710,"_405":523,"_406":35,"_216":635},{"_116":710,"_405":623,"_406":35,"_216":635},{"_116":739,"_405":520,"_406":35,"_216":740},{"_116":727,"_405":526,"_406":92,"_216":728},{"_116":737,"_405":520,"_406":35,"_216":738},{"_116":727,"_405":523,"_406":92,"_216":728},{"_116":735,"_405":622,"_406":35,"_216":736},{"_116":727,"_405":644,"_406":92,"_216":728},{"_116":707,"_405":523,"_406":35,"_216":709},{"_116":733,"_405":520,"_406":92,"_216":734},{"_116":731,"_405":520,"_406":35,"_216":732},{"_116":729,"_405":520,"_406":92,"_216":730},{"_116":727,"_405":618,"_406":92,"_216":728},{"_116":725,"_405":520,"_406":35,"_216":726},{"_116":723,"_405":520,"_406":35,"_216":724},{"_116":721,"_405":520,"_406":92,"_216":722},{"_116":719,"_405":627,"_406":35,"_216":720},{"_116":707,"_405":618,"_406":35,"_216":709},{"_116":717,"_405":520,"_406":35,"_216":718},{"_116":715,"_405":520,"_406":92,"_216":716},{"_116":713,"_405":520,"_406":35,"_216":714},{"_116":711,"_405":520,"_406":35,"_216":712},{"_116":710,"_405":520,"_406":35,"_216":635},{"_116":707,"_405":708,"_406":35,"_216":709},{"_116":705,"_405":520,"_406":35,"_216":706},{"_116":703,"_405":520,"_406":92,"_216":704},{"_116":701,"_405":520,"_406":35,"_216":702},"_ga_xxxxxxxxxx","Used by Google Analytics to identify and track an individual session with your device.","s_ppvl","This cookie is set by Adobe Analytics to store information on the percentage of the page displayed to the end user.","coveo_visitorId","This cookie name is associated with Coveo which is used to store a unique user ID for marketing purposes and to track the number of distinct users who submitted a query or clicked an item, counting them only once even if they did the same query or clicked the same item over the course of many visits.","_dd_s","marketplace.nvidia.com","Datadog Cookie used to group all events generated from a unique user session across multiple pages. It contains the current session ID, whether the session is excluded due to sampling, and the expiration date of the session. The cookie is extended for an extra 15 minutes every time the user interacts with the website, up to the maximum user session duration (4 hours).","nvweb_A","_ga","This cookie name is associated with Google Universal Analytics - which is a significant update to Google's more commonly used analytics service. This cookie is used to distinguish unique users by assigning a randomly generated number as a client identifier. It is included in each page request in a site and used to calculate visitor, session and campaign data for the sites analytics reports. By default it is set to expire after 2 years, although this is customisable by website owners.","_uetvid","This is a cookie utilised by Microsoft Bing Ads and is a tracking cookie. It allows us to engage with a user that has previously visited our website.","tp","This cookie is associated with Adobe Analytics site tracking. It registers if the website's scroll-depth detection is active. This function remembers how far the user has scrolled on the current session across the website's sub-pages.","_hly_vid","This cookie _hly_vid is used to record website and server performance with each visit to the site. The data is aggregated and analyzed; therefore, it remains anonymous upon use.","_session_id","This cookie creates an interim session ID used as an in-session user ID.\n","s_cc","This cookie is associated with the Adobe Site Catalyst. It determines whether cookies are enabled in the web browser.","s_ecid","Adobe Analytics cookie containing the the Adobe Experience Cloud 1st party visitor id if the Analytics AMCV visitor id is not found.","_hjSessionUser_xxxxxx","Hotjar cookie that is set when a user first lands on a page with the Hotjar script. It is used to persist the Hotjar User ID, unique to that site on the browser. This ensures that behavior in subsequent visits to the same site will be attributed to the same user ID.","_hjHasCachedUserAttributes","This cookie name is associated with Hotjar which is used to know whether the data set in _hjUserAttributes Local Storage item is up to date or not.","s_ppv","This cookie is associated with Adobe Analytics site tracking and is used to measure a user's scroll activity to see how much of a page they view before moving to another page.","_gid","This cookie name is associated with Google Universal Analytics. This appears to be a new cookie and as of Spring 2017 no information is available from Google. It appears to store and update a unique value for each page visited.","_hjTLDTest","When the Hotjar script executes we try to determine the most generic cookie path we should use, instead of the page hostname. This is done so that cookies can be shared across subdomains (where applicable). To determine this, we try to store the _hjTLDTest cookie for different URL substring alternatives until it fails. After this check, the cookie is removed.","ekfls-2025-03-27","This cookie name is associated with readme.io which is used to help maintain data for the API reference support page.","s_getNewRepeat","This cookie is used to determine whether a visitor is a new visitor or a repeat visitor to the site, and is used for analytical purposes.","_hly_sid","This cookie is likely associated with website analytics, as the prefix '_hly_' suggests a connection to Holistics, a business intelligence and analytics platform.","docs.nvidia.com","_hjSession_xxxxxx","A cookie that holds the current session data. This ensues that subsequent requests within the session window will be attributed to the same Hotjar session.","C0001","Required Cookies","These cookies enable core functionality such as security, network management, and accessibility. These cookies are required for the site to function and cannot be turned off.",[889,890,891,892,893,894,895,896,897,898,899,900,901,902,903,904,905,906,907,908,909,910,911,912,913,914,915,916,917,918,919,920,921,922,923,924,925,926,927,928,929,930,931,932,933,934,935,936,937,938,939,940,941,942,943,944,945,946,947,948,949,950,951,952,953,954,955,956,957,958,959,960,961],[749,750,751,752,753,754,755,756,757,758,759,760,761,762,763,764,765,766,767,768,769,770,771,772,773],{"_398":886,"_251":886,"_216":400,"_371":887},{"_398":880,"_251":880,"_216":400,"_371":881},{"_398":877,"_251":877,"_216":400,"_371":878},{"_398":865,"_251":865,"_216":400,"_371":866},{"_398":862,"_251":862,"_216":400,"_371":863},{"_398":859,"_251":859,"_216":400,"_371":860},{"_398":854,"_251":854,"_216":400,"_371":855},{"_398":461,"_251":461,"_216":400,"_371":851},{"_398":848,"_251":848,"_216":400,"_371":849},{"_398":453,"_251":453,"_216":400,"_371":846},{"_398":448,"_251":448,"_216":400,"_371":844},{"_398":839,"_251":839,"_216":400,"_371":840},{"_398":834,"_251":834,"_216":400,"_371":835},{"_398":829,"_251":829,"_216":400,"_371":830},{"_398":826,"_251":826,"_216":400,"_371":827},{"_398":574,"_251":574,"_216":400,"_371":823},{"_398":433,"_251":433,"_216":400,"_371":820},{"_398":569,"_251":569,"_216":400,"_371":813},{"_398":565,"_251":565,"_216":400,"_371":807},{"_398":801,"_251":801,"_216":400,"_371":802},{"_398":796,"_251":796,"_216":400,"_371":797},{"_398":791,"_251":791,"_216":400,"_371":792},{"_398":787,"_251":787,"_216":400,"_371":788},{"_398":782,"_251":782,"_216":400,"_371":783},{"_398":774,"_251":774,"_216":400,"_371":775},"polyfill.io",[776,777],{"_116":780,"_405":774,"_406":35,"_216":781},{"_116":778,"_405":774,"_406":35,"_216":779},"cf_ob_info","This cookie is associated with sites using CloudFlare. It is used to improve page load times and to override any security restrictions based on the visitors IP address. It does not contain any user identification information.","cf_use_ob","This cookie is associated with sites using CloudFlare. It is used to improve page load times and to override any security restrictions based on the visitors IP address. It does not contain any user identification information. ","www.recaptcha.net",[784],{"_116":785,"_405":782,"_406":35,"_216":786},"_GRECAPTCHA","This domain is used by Google reCAPTCHA, which provides spam protection services to websites by detecting if a user is a human or a bot.","api2.hcaptcha.com",[789],{"_116":790,"_405":787,"_406":35,"_216":442},"__cflb","instagram.com",[793],{"_116":794,"_405":791,"_406":35,"_216":795},"csrftoken","Instagram is a social networking service owned by Facebook, Inc. It's designed for sharing photos and videos from a smartphone. The service also uses cookies for various purposes, including analytics, personalization, and advertising.","readme.io",[798],{"_116":799,"_405":796,"_406":35,"_216":800},"__cf_bm","The __cf_bm cookie is a cookie necessary to support Cloudflare Bot Management, currently in private beta. As part of our bot management service, this cookie helps manage incoming traffic that matches criteria associated with bots.","nvidiacorp.us-5.evergage.com",[803,804],{"_116":806,"_405":801,"_406":35,"_216":442},{"_116":805,"_405":801,"_406":35,"_216":442},"AWSALBTG","AWSALBTGCORS",[808,809],{"_116":811,"_405":565,"_406":35,"_216":812},{"_116":810,"_405":565,"_406":35,"_216":442},"ak_bmsc","bm_mi","This domain is likely associated with NVIDIA's partner program, which could involve marketing, sales, and support services for partners.",[814,815,816],{"_116":819,"_405":569,"_406":35,"_216":442},{"_116":818,"_405":569,"_406":92,"_216":442},{"_116":817,"_405":569,"_406":35,"_216":442},"AWSALBCORS","JSESSIONID","AWSALB",[821],{"_116":799,"_405":433,"_406":35,"_216":822},"This is a CloudFoundry cookie",[824],{"_116":825,"_405":574,"_406":35,"_216":578},"logged_in","hcaptcha.com",[828],{"_116":799,"_405":826,"_406":35,"_216":822},"tracking.intentsify.io",[831],{"_116":832,"_405":829,"_406":35,"_216":833},"userId","Used to track usage of the embedded video player.","static.reo.dev",[836],{"_116":837,"_405":834,"_406":35,"_216":838},"cookietest","Common cookie name could have a number of different origins. Where this is first party and a session cookie, its most likely to do with checking to see if the browser is set to block or allow cookies.","www.datadoghq-browser-agent.com",[841],{"_116":842,"_405":839,"_406":35,"_216":843},"dd_cookie_test_","dd_cookie_test",[845],{"_116":799,"_405":448,"_406":35,"_216":822},[847],{"_116":799,"_405":453,"_406":35,"_216":822},"supabase.co",[850],{"_116":799,"_405":848,"_406":35,"_216":822},[852],{"_116":853,"_405":461,"_406":35,"_216":468},"VISITOR_PRIVACY_METADATA","dosa-prd.azurewebsites.net",[856],{"_116":857,"_405":854,"_406":35,"_216":858},"access_token","This cookie is used to obtain the permissions for users of the portal to allow for API access in order to retrieve information required to perform portal functions.","w.hcaptcha.com",[861],{"_116":799,"_405":859,"_406":35,"_216":822},"cdn.readme.io",[864],{"_116":799,"_405":862,"_406":35,"_216":822},"queue-it.net",[867,868,869,870,871],{"_116":876,"_405":865,"_406":35,"_216":400},{"_116":875,"_405":865,"_406":35,"_216":400},{"_116":874,"_405":865,"_406":35,"_216":400},{"_116":873,"_405":865,"_406":35,"_216":400},{"_116":872,"_405":865,"_406":35,"_216":400},"Queue-it-{queueId}","Queue-it-{dmonvidia }________{eventId}","Queue-it-token","queueitsoftblock_{queueId}","Queue-it-{nvidia}________{eventId}","x.com",[879],{"_116":799,"_405":877,"_406":35,"_216":800},"api.hcaptcha.com",[882,883],{"_116":885,"_405":880,"_406":35,"_216":442},{"_116":790,"_405":880,"_406":35,"_216":884},"This domain is associated with hCaptcha, a service that provides CAPTCHA and reCAPTCHA alternatives for user verification and bot protection.","hmt_id","vimeo.com",[888],{"_116":799,"_405":886,"_406":35,"_216":822},{"_116":1022,"_405":520,"_406":35,"_216":1053},{"_116":1051,"_405":615,"_406":92,"_216":1052},{"_116":1044,"_405":1013,"_406":92,"_216":1045},{"_116":1050,"_405":708,"_406":92,"_216":400},{"_116":1048,"_405":520,"_406":35,"_216":1049},{"_116":806,"_405":1034,"_406":35,"_216":973},{"_116":981,"_405":1043,"_406":92,"_216":983},{"_116":805,"_405":972,"_406":35,"_216":973},{"_116":984,"_405":1043,"_406":35,"_216":985},{"_116":1046,"_405":520,"_406":92,"_216":1047},{"_116":1044,"_405":1020,"_406":92,"_216":1045},{"_116":978,"_405":1043,"_406":35,"_216":980},{"_116":1041,"_405":520,"_406":35,"_216":1042},{"_116":805,"_405":964,"_406":35,"_216":973},{"_116":969,"_405":1040,"_406":35,"_216":971},{"_116":817,"_405":970,"_406":35,"_216":986},{"_116":819,"_405":970,"_406":35,"_216":1037},{"_116":984,"_405":979,"_406":35,"_216":985},{"_116":1038,"_405":618,"_406":35,"_216":1039},{"_116":819,"_405":644,"_406":35,"_216":1037},{"_116":1035,"_405":520,"_406":35,"_216":1036},{"_116":811,"_405":644,"_406":35,"_216":1030},{"_116":1007,"_405":623,"_406":92,"_216":400},{"_116":805,"_405":1034,"_406":35,"_216":973},{"_116":1032,"_405":972,"_406":92,"_216":1033},{"_116":996,"_405":520,"_406":35,"_216":987},{"_116":969,"_405":520,"_406":35,"_216":971},{"_116":1031,"_405":532,"_406":35,"_216":400},{"_116":811,"_405":520,"_406":35,"_216":1030},{"_116":819,"_405":964,"_406":35,"_216":1029},{"_116":1027,"_405":627,"_406":92,"_216":1028},{"_116":1025,"_405":520,"_406":92,"_216":1026},{"_116":1024,"_405":520,"_406":35,"_216":400},{"_116":1022,"_405":970,"_406":35,"_216":1023},{"_116":1021,"_405":532,"_406":35,"_216":400},{"_116":1012,"_405":1020,"_406":92,"_216":1014},{"_116":1018,"_405":623,"_406":35,"_216":1019},{"_116":1016,"_405":708,"_406":35,"_216":1017},{"_116":1015,"_405":526,"_406":35,"_216":400},{"_116":1012,"_405":1013,"_406":92,"_216":1014},{"_116":1010,"_405":979,"_406":92,"_216":1011},{"_116":1008,"_405":520,"_406":35,"_216":1009},{"_116":1007,"_405":523,"_406":92,"_216":400},{"_116":978,"_405":982,"_406":35,"_216":980},{"_116":810,"_405":997,"_406":35,"_216":987},{"_116":1005,"_405":523,"_406":35,"_216":1006},{"_116":1003,"_405":520,"_406":35,"_216":1004},{"_116":1001,"_405":520,"_406":35,"_216":1002},{"_116":1000,"_405":520,"_406":35,"_216":400},{"_116":842,"_405":523,"_406":35,"_216":999},{"_116":996,"_405":644,"_406":35,"_216":987},{"_116":799,"_405":615,"_406":35,"_216":998},{"_116":996,"_405":997,"_406":35,"_216":987},{"_116":995,"_405":532,"_406":35,"_216":400},{"_116":994,"_405":520,"_406":35,"_216":967},{"_116":992,"_405":979,"_406":35,"_216":993},{"_116":805,"_405":990,"_406":35,"_216":991},{"_116":989,"_405":532,"_406":35,"_216":400},{"_116":988,"_405":520,"_406":35,"_216":400},{"_116":981,"_405":979,"_406":92,"_216":983},{"_116":810,"_405":644,"_406":35,"_216":987},{"_116":817,"_405":644,"_406":35,"_216":986},{"_116":984,"_405":982,"_406":35,"_216":985},{"_116":981,"_405":982,"_406":92,"_216":983},{"_116":978,"_405":979,"_406":35,"_216":980},{"_116":976,"_405":972,"_406":92,"_216":977},{"_116":974,"_405":630,"_406":35,"_216":975},{"_116":806,"_405":972,"_406":35,"_216":973},{"_116":969,"_405":970,"_406":35,"_216":971},{"_116":799,"_405":622,"_406":35,"_216":968},{"_116":966,"_405":520,"_406":35,"_216":967},{"_116":817,"_405":964,"_406":35,"_216":965},{"_116":962,"_405":520,"_406":35,"_216":963},"LoginHints","Keeps users logged in across NVIDIA domains.","api.ngc.nvidia.com","This cookie is managed by AWS and is used for load balancing.","bm_so","This cookie name is used to load the site as quickly as possible and to recognize bots using Akamai.","The __cf_bm cookie is a cookie necessary to support Cloudflare Bot Management, currently in private beta. As part of our bot management service, this cookie helps manage incoming traffic that matches criteria associated with bots.\nThis is a CloudFoundry cookie","OptanonConsent","nvid.nvidia.com","This cookie is set by the cookie compliance solution from OneTrust. It stores information about the categories of cookies the site uses and whether visitors have given or withdrawn consent for the use of each category. This enables site owners to prevent cookies in each category from being set in the users browser, when consent is not given. The cookie has a normal lifespan of one year, so that returning visitors to the site will have their preferences remembered. It contains no information that can identify the site visitor.","events.nvidia.com","This cookie is associated with the Amazon Web Services Elastic Load Balancing functionality for routing client request on the server.Used to attribute commission to affiliates when you arrive at the website from an affiliate referral link. It is set when you click on one of our links and used to let the advertiser and us know the website from which you came.","sessionid","This is a very generic cookie name that may have different purposes on different sites, but generally it will be some kind of anonymous session identifier.","_csrf","This cookie is used to prevent Cross-Site Request Forgery (CSRF) attacks, ensuring that each request made to the server is from the user's browser and is intentional.","LSKey-c$CookieConsentPolicy","enterprise-support.nvidia.com","Used to apply end-user cookie consent preferences set by our client-side utility.Custom- Short Duration. Unique ID of an Enquiry- used to validate access to case, and load exisiting data.\n","renderCtx","academicgrants.nvidia.com","Used to store site parameters in the session for reuse across requests by a single client for functionality and performance reasons.","CookieConsentPolicy","Used to apply end-user cookie consent preferences set by our client-side utility.","This cookie is managed by AWS and is used for load balancing.Used to attribute commission to affiliates when you arrive at the website from an affiliate referral link. It is set when you click on one of our links and used to let the advertiser and us know the website from which you came.","This cookie is associated with Akamai and is used to differentiate between traffic from humans and bots.","nv_co","__uxq412__ses.88c1","learn.learn.nvidia.com","This cookie is associated with the Amazon Web Services Elastic Load Balancing functionality for routing client request on the server.","pctrk","Used to count page views by unauthenticated users against license usage.","bm_s","bpazaws52gukakzc__ctrl0_ctl36_uccaptcha","bm_sv","store.nvidia.com","The __cf_bm cookie is a cookie necessary to support Cloudflare Bot Management, currently in private beta. As part of our bot management service, this cookie helps manage incoming traffic that matches criteria associated with bots.\n","This cookie name is used by DataDog to determine if cookies are supported on the website.dd_cookie_test","AMP_TLDTEST_xxxxxxxx","SESSlsd","This cookie is likely to be a session identifier used to maintain user session state across page requests.","mboxEdgeCluster","This cookie name is associated with Adobe Target which is used to communicate the correct edge cluster to at.js so that Target profiles can stay in sync as users navigate across a site.","_rdt_em","This cookie name is associated with Reddit which is used to manage session state and user interactions.","megaNavCollapsed","_abck","The _abck cookie is associated with Akamai's Bot Manager, which is used to prevent fraud and abuse by safeguarding online interactions and detecting irregularities that could suggest malicious activity, such as bot-driven attacks or hacking attempts.","idccsrf","This cookie name is associated with Salesforce which is used to track CrossSiteRequestForgery validation for certain single sign-on (SSO) flows.","ApplicationGatewayAffinity","partners.nvidia.com","This cookie is used to maintain a user session by the server in a load balanced environment, to ensure the user has a consistent experience.","AMP_TEST_xxxxxxxx","bm_lso","This cookie name is associated with Akamai which is used to load the site as quickly as possible and to recognize bots.","SSESS###########","This cookie is likely to be a session cookie used to maintain an anonymized user session by the server in a secure environment, often seen in sites built with the Drupal content management system.","nvdeveloper.nvidia.com","__uxq412__id.88c1","OptanonAlertBoxClosed","This cookie is set by websites using certain versions of the cookie law compliance solution from OneTrust. It is set after visitors have seen a cookie information notice and in some cases only when they actively close the notice down. It enables the website not to show the message more than once to a user. The cookie has a normal lifespan of one year and contains no personal information.","QueueITAccepted-SDFrts345E-V3_{eventId}","TEST_AMCV_COOKIE_WRITE","This cookie is used to test whether cookies are enabled in the browser. It is a temporary cookie used to determine if the browser accepts cookies.","XSRF-TOKEN","This cookie is written to help with site security in preventing Cross-Site Request Forgery attacks.This cookie helps with site security in preventing Cross-Site Request Forgery attacks.","AWS ELB application load balancer","This cookie is associated with Akamai. It is used to analyse traffic to determine if it is automated traffic generated by bots or a human user.","bpazaws52gukakzc__ctrl0_ctl78_list_ctl09_systemcaptcha","PHPSESSID","PHP session cookie associated with embedded content from this domain.\n","base.manager.learn.nvidia.com","bm_ss","This cookie is used to help protect websites from automated bot traffic.","AWS ELB application load balancerUsed to attribute commission to affiliates when you arrive at the website from an affiliate referral link. It is set when you click on one of our links and used to let the advertiser and us know the website from which you came.","_devzone_session","This cookie is likely used to maintain a user session on the website by the server. It is common for session identifiers to be used to keep users logged in and track their interaction with the site during a session.","docs.ngc.nvidia.com","bm_sz","This cookie is used by Akamai to optimize site performance and security. It is typically associated with the delivery of web content and services.","vcalliance.nvidia.com","ASP.NET_SessionId","General purpose platform session cookie, used by sites written with Microsoft .NET based technologies. Usually used to maintain an anonymised user session by the server.","com.adobe.alloy.getTld","This cookie name is associated with the Adobe WebSDK which is used to determine the top-most domain that can accept cookies.","SESSls","This cookie is likely to be a session identifier, used to maintain a user's session state across page requests.","c_code","BIGipServerx","This cookie name is associated with the BIG-IP product suite from company F5. Usually associated with managing sessions on load balanced servers, to ensure user requests are routed consistently to the correct server. The common root is BIGipServer most commonly followed by a domain name, usually the one that it is hosted on, but not always.","This cookie is set by websites using certain versions of the cookie law compliance solution from OneTrust. It is set after visitors have seen a cookie information notice and in some cases only when they actively close the notice down. It enables the website not to show the message more than once to a user. The cookie has a one year lifespan and contains no personal information.","default","

NVIDIA uses cookies to improve your experience on our web site. We and our third-party partners also use cookies and other tools to collect and record information you provide as well as information about your interactions with our websites for performance improvement, analytics, and to assist in marketing efforts. By clicking \"Accept All\", you consent to our use of cookies and other tools as described in our Cookie Policy. You can manage your cookie settings by clicking on \"Manage Settings.\" By continuing to use this site or by clicking one of the buttons below, you agree to our Terms of Service (which contains important waivers). Please see our Privacy Policy for more information on our privacy practices.

","gpc","

We have detected the Global Privacy Control (GPC) signal and recorded your rejection of all optional cookies on this site for this browser. You can manage your cookie settings by clicking on \"Manage Settings\". Please see our Cookie Policy for more information. To opt out of non-cookie personal information \"sales\" / \"sharing\" for targeted advertising purposes, please visit the NVIDIA Preference Center. Please see our Privacy Policy for more information on our privacy practices.

","gpcLoggedIn","

We have detected the Global Privacy Control Signal (GPC) and have opted you out of all optional cookies on this browser. You can manage your cookie settings by clicking on \"Manage Settings\". Please see our Cookie Policy for more information. We have also opted you out of \"sharing\"/\"sales\" of personal information outside of cookies. You can manage these settings in the NVIDIA NVIDIA Preference Center. Please see our Privacy Policy for more information.

","gpcLoggedInOverridden","

We have detected the Global Privacy Control Signal (GPC) and have opted you out of all optional cookies on this browser. You can manage your cookie settings by clicking on \"Manage Settings\". Please see our Cookie Policy for more information. We have also opted you out of \"sharing\"/\"sales\" of personal information outside of cookies which overrides at least one of your previous settings. You can manage them in the NVIDIA Preference Center. Please see our Privacy Policy for more information.

",{"_744":18,"_647":60,"_555":60,"_375":60,"_365":60},"hasConsented","gpcApplied","showBanner","geolocation",{"_1068":1069,"_1070":1071},"country","US","state","OR","query","categories",[1091,1092,1093,1094,1095],"optionUrls",[1079,1080,1081,1082,1083,1084,1085,1086,1087,1088,1089,1090],"total",12,"/orgs/nim/deepseek-ai/containers/deepseek-r1/-","/orgs/nim/nvidia/containers/llama-3.1-nemotron-70b-instruct/-","/orgs/nvidia/-/containers/pytorch/-","/orgs/nvidia/omniverse/collections/kit/-","/orgs/nvidia/-/collections/deepstream_sdk/-","/orgs/nvidia/omniverse/collections/kit-appstreaming-collection/-","/orgs/nvidia/research/models/stylegan3/-","/orgs/nvidia/tao/models/peoplenet/-","/orgs/nvidia/tao/models/trafficcamnet/-","/orgs/nvidia/riva/resources/riva_quickstart/-","/orgs/nvidia/-/helm-charts/gpu-operator/-","/orgs/nvidia/-/helm-charts/k8s-nim-operator/-",{"_21":1169,"_1097":1170,"_1099":60},{"_21":1134,"_1097":1135,"_1099":63},{"_21":1117,"_1097":1118,"_1099":1119},{"_21":22,"_1097":1110,"_1099":1111},{"_21":1096,"_1097":1098,"_1099":20},"helm-charts","items",[1100,1101],"startIndex",{"_21":1096,"_251":1107,"_210":211,"_212":-7,"_246":1108,"_254":1104,"_279":1109,"_1106":1089},{"_21":1096,"_251":1102,"_210":211,"_212":-7,"_246":1103,"_254":1104,"_279":1105,"_1106":1090},"NVIDIA NIM Operator","k8s-nim-operator","NVIDIA",[],"url","GPU Operator","gpu-operator",[],[1112],9,{"_21":22,"_251":1113,"_210":211,"_212":1114,"_246":1115,"_254":1104,"_279":1116,"_1106":1088},"Riva Skills Quick Start","riva","riva_quickstart",[],"models",[1120,1121,1122],6,{"_21":1117,"_251":1130,"_210":211,"_212":1131,"_246":1132,"_254":1104,"_279":1133,"_1106":1085},{"_21":1117,"_251":1127,"_210":211,"_212":1124,"_246":1128,"_254":1104,"_279":1129,"_1106":1086},{"_21":1117,"_251":1123,"_210":211,"_212":1124,"_246":1125,"_254":1104,"_279":1126,"_1106":1087},"TrafficCamNet","tao","trafficcamnet",[],"PeopleNet","peoplenet",[],"StyleGAN3 pretrained models","research","stylegan3",[],"collections",[1136,1137,1138],{"_21":1134,"_251":1166,"_210":211,"_212":1140,"_246":1167,"_254":1104,"_279":1168,"_1106":1082},{"_21":1134,"_251":1163,"_210":211,"_212":-7,"_246":1164,"_254":1104,"_279":1165,"_1106":1083},{"_21":1134,"_251":1139,"_210":211,"_212":1140,"_246":1141,"_254":1104,"_279":1142,"_1106":1084},"Omniverse Kit App Streaming","omniverse","kit-appstreaming-collection",[1143,1144,1145],{"_116":1157,"_251":1158,"_248":1159,"_258":1160,"_1150":1161},{"_116":1140,"_251":1153,"_248":1154,"_258":1149,"_1150":1155},{"_116":1146,"_251":1147,"_248":1148,"_258":1149,"_1150":1151},"omniverse-dev","NVIDIA Developer Program","Join the Developer Program for access to free tools, support, and tech resources.","https://store-images.s-microsoft.com/image/apps.33936.6f1bcf42-55d0-45d1-b4b6-f08c68a50bbe.e91ca707-db64-49d6-9761-4781590a9685.7ba64fc5-7c44-4864-a5e6-7bde05126935","supportedFeatures",[1152],"UNIFIED_CATALOG","NVIDIA Omniverse","NVIDIA Omniverse is a platform of APIs, services, and software development kits (SDKs) that enable developers to build generative AI-enabled tools and applications.",[1152,1156],"SHAREABLE_TO_USER_GROUPS","nv-ai-enterprise","NVIDIA AI Enterprise","Accelerate your AI agent development","https://assets.ngc.nvidia.com/products/catalog/images/product-NVAIE.jpg",[1152,1156,1162],"NGC_CONSOLE","DeepStream SDK","deepstream_sdk",[],"Omniverse Kit (FB)","kit",[1143,1144,1145],"containers",[1171,1172,1173],{"_21":1169,"_251":1185,"_210":1178,"_212":1186,"_246":1187,"_254":1104,"_279":1188,"_1106":1079},{"_21":1169,"_251":1177,"_210":1178,"_212":211,"_246":1179,"_254":1104,"_279":1180,"_1106":1080},{"_21":1169,"_251":1174,"_210":211,"_212":-7,"_246":1175,"_254":1104,"_279":1176,"_1106":1081},"PyTorch","pytorch",[],"Llama-3.1-Nemotron-70B-Instruct","nim","llama-3.1-nemotron-70b-instruct",[1143,1181],{"_116":1182,"_251":1147,"_248":1148,"_258":1183,"_1150":1184},"nim-dev","https://assets.ngc.nvidia.com/products/catalog/images/product-nim-dev.jpg",[1152],"DeepSeek-R1","deepseek-ai","deepseek-r1",[1143,1181],"ownerOrgs",[],"memberOrgs",[],"externalOrgs",[],"apiKey","2b44f86fea50e90ffd2db29ee2028419","consentGranted","applicationId","4c857e4f-5ffa-45fa-b75e-af773c22f0ad","clientToken","pubb6927db33dae60ba5031a186cde0242a","site","datadoghq.com","service","catalog-ui","env","sessionSampleRate",100,"sessionReplaySampleRate","compressIntakeRequests","storeContextsAcrossPages","defaultPrivacyLevel","mask-user-input","trackViewsManually","trackingConsent","not-granted",{"_1218":1219,"_1220":312,"_1221":35,"_1222":1223,"_1097":1224,"_1225":1226,"_1227":1219,"_1228":35,"_1229":35,"_1230":35,"_1231":1232},"appName","ngc-catalog","appOrigin","hideOnPrivateOrg","icon","ObjectsModules2",[1233,1234,1235,1236,1237,1238,1239],"menuIcon","editor-swatches","productName","requiresAllScopes","requiresLogin","requiresOrgLevelScopesOnly","title","Catalog",{"_1231":1261,"_51":1262,"_1225":1263},{"_1254":1255,"_1231":1256,"_51":1257,"_1227":1157},{"_1231":1252,"_51":1253,"_1225":1226},{"_1231":1249,"_51":1250,"_1225":1251},{"_1231":1246,"_51":1247,"_1225":1248},{"_1231":1243,"_51":1244,"_1225":1245},{"_1231":1240,"_51":1241,"_1225":1242},"Resources","/resources","misc-library","Models","/models","misc-neural-network","Helm Charts","/helm-charts","misc-ship-wheel","Containers","/containers","3d-cube","Collections","/collections","access",[1258],"Enterprise Catalog","/enterprise",[1259,1260],"NVAIE","READ","Explore Catalog","/","common-home",{"_163":164,"_165":166,"_167":168,"_169":170,"_171":172,"_173":174},{"_162":35}]

	Files	Compressed Size		Download
22.11.0Selected	1	43.72 KB	December 5, 2022 UTC
22.04.0	1	43.03 KB	September 22, 2022 UTC