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

This repository provides an implementation of the Jasper model in PyTorch from the paper Jasper: An End-to-End Convolutional Neural Acoustic Model https://arxiv.org/pdf/1904.03288.pdf.\nThe Jasper model is an end-to-end neural acoustic model for automatic speech recognition (ASR) that provides near state-of-the-art results on LibriSpeech among end-to-end ASR models without any external data. The Jasper architecture of convolutional layers was designed to facilitate fast GPU inference, by allowing whole sub-blocks to be fused into a single GPU kernel. This is important for meeting strict real-time requirements of ASR systems in deployment.

The results of the acoustic model are combined with the results of external language models to get the top-ranked word sequences\ncorresponding to a given audio segment. This post-processing step is called decoding.

This repository is a PyTorch implementation of Jasper and provides scripts to train the Jasper 10x5 model with dense residuals from scratch on the Librispeech dataset to achieve the greedy decoding results of the original paper.\nThe original reference code provides Jasper as part of a research toolkit in TensorFlow openseq2seq.\nThis repository provides a simple implementation of Jasper with scripts for training and replicating the Jasper paper results.\nThis includes data preparation scripts, training and inference scripts.\nBoth training and inference scripts offer the option to use Automatic Mixed Precision (AMP) to benefit from Tensor Cores for better performance.

In addition to providing the hyperparameters for training a model checkpoint, we publish a thorough inference analysis across different NVIDIA GPU platforms, for example, DGX A100, DGX-1, DGX-2 and T4.

This model is trained with mixed precision using Tensor Cores on Volta, Turing, and the NVIDIA Ampere GPU architectures. Therefore, researchers can get results 3x 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.

The original paper takes the output of the Jasper acoustic model and shows results for 3 different decoding variations: greedy decoding, beam search with a 6-gram language model and beam search with further rescoring of the best ranked hypotheses with Transformer XL, which is a neural language model. Beam search and the rescoring with the neural language model scores are run on CPU and result in better word error rates compared to greedy decoding.\nThis repository provides instructions to reproduce greedy decoding results. To run beam search or rescoring with TransformerXL, use the following scripts from the openseq2seq repository:\nhttps://github.com/NVIDIA/OpenSeq2Seq/blob/master/scripts/decode.py\nhttps://github.com/NVIDIA/OpenSeq2Seq/tree/master/external_lm_rescore

Model architecture

Details on the model architecture can be found in the paper Jasper: An End-to-End Convolutional Neural Acoustic Model.

\n\n\n\n\n\n\n\n\n\n\n\n\n\n

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


Figure 1: Jasper BxR model: B- number of blocks, R- number of sub-blocks	Figure 2: Jasper Dense Residual

Jasper is an end-to-end neural acoustic model that is based on convolutions.\nIn the audio processing stage, each frame is transformed into mel-scale spectrogram features, which the acoustic model takes as input and outputs a probability distribution over the vocabulary for each frame.\nThe acoustic model has a modular block structure and can be parametrized accordingly:\na Jasper BxR model has B blocks, each consisting of R repeating sub-blocks.

Each sub-block applies the following operations in sequence: 1D-Convolution, Batch Normalization, ReLU activation, and Dropout.

Each block input is connected directly to the last subblock of all following blocks via a residual connection, which is referred to as dense residual in the paper.\nEvery block differs in kernel size and number of filters, which are increasing in size from the bottom to the top layers.\nIrrespective of the exact block configuration parameters B and R, every Jasper model has four additional convolutional blocks:\none immediately succeeding the input layer (Prologue) and three at the end of the B blocks (Epilogue).

The Prologue is to decimate the audio signal\nin time in order to process a shorter time sequence for efficiency. The Epilogue with dilation captures a bigger context around an audio time step, which decreases the model word error rate (WER).\nThe paper achieves best results with Jasper 10x5 with dense residual connections, which is also the focus of this repository and is in the following referred to as Jasper Large.

Default configuration

The following features were implemented in this model:

GPU-supported feature extraction with data augmentation options SpecAugment and Cutout
offline and online Speed Perturbation
data-parallel multi-GPU training and evaluation
AMP with dynamic loss scaling for Tensor Core training
FP16 inference

Competitive training results and analysis is provided for the following Jasper model configuration

\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\">

Model	Number of Blocks	Number of Subblocks	Max sequence length	Number of Parameters
Jasper Large	10	5	16.7 s	333 M

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

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

Feature	Jasper
Apex AMP	Yes
Apex DistributedDataParallel	Yes

Features

Apex AMP - a tool that enables Tensor Core-accelerated training. Refer to the Enabling mixed precision section for more details.

Apex\nDistributedDataParallel -\na module wrapper that enables easy multiprocess distributed data parallel\ntraining, similar to\ntorch.nn.parallel.DistributedDataParallel.\nDistributedDataParallel is optimized for use with\nNCCL. It achieves high performance by\noverlapping communication with computation during backward() and bucketing\nsmaller gradient transfers to reduce the total number of transfers required.

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 half-precision format, while storing minimal information in single-precision to retain 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 overall speedup on the most arithmetically 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.

The ability to train deep learning networks with lower precision was introduced in the Pascal architecture and first supported in CUDA 8 in the NVIDIA Deep Learning SDK.

For information about:

How to train using mixed precision, see theMixed 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.
APEX tools for mixed precision training, see the NVIDIA Apex: Tools for Easy Mixed-Precision Training in PyTorch.

Enabling mixed precision

For training, mixed precision can be enabled by setting the flag: train.py --amp. When using bash helper scripts: scripts/train.sh scripts/inference.sh, etc., mixed precision can be enabled with env variable AMP=true.

Mixed precision is enabled in PyTorch by using the Automatic Mixed Precision\n(AMP) library from APEX that casts variables\nto half-precision upon retrieval, while storing variables in single-precision\nformat. Furthermore, to preserve small gradient magnitudes in backpropagation,\na loss\nscaling\nstep must be included when applying gradients. In PyTorch, loss scaling can be\neasily applied by using scale_loss() method provided by AMP. The scaling\nvalue to be used can be\ndynamic or fixed.

For an in-depth walk through on AMP, check out sample usage\nhere. APEX is a PyTorch extension that contains\nutility libraries, such as AMP, which require minimal network code changes to\nleverage Tensor Cores performance.

The following steps were needed to enable mixed precision training in Jasper:

Import AMP from APEX (file: train.py):

from apex import amp\n

Initialize AMP and wrap the model and the optimizer

   model, optimizer = amp.initialize(\n     min_loss_scale=1.0,\n     models=model,\n     optimizers=optimizer,\n     opt_level='O1')\n\n

Apply scale_loss context manager

with amp.scale_loss(loss, optimizer) as scaled_loss:\n    scaled_loss.backward()\n

Enabling 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 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 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

Acoustic model\nAssigns a probability distribution over a vocabulary of characters given an audio frame.

Language Model\nAssigns a probability distribution over a sequence of words. Given a sequence of words, it assigns a probability to the whole sequence.

Pre-training\nTraining a model on vast amounts of data on the same (or different) task to build general understandings.

Automatic Speech Recognition (ASR)\nUses both acoustic model and language model to output the transcript of an input audio signal.

","resourceVersions",[232,233],"resourceVersionsTotal","hasMoreVersions","versionMetadata",{"_211":28,"_212":-5,"_52":117,"_114":-5,"_118":213,"_120":-5,"_214":116,"_125":16,"_122":34,"_215":-5,"_121":34,"_216":217,"_218":91,"_219":220,"_221":91,"_222":223,"_224":91,"_225":226,"_227":91,"_228":229,"_230":91,"_123":231},"versionNotFound","isLoggedIn","latestVersionPath","/orgs/nvidia/dle/resources/jasper_pyt/-","-","imageReference","cliDownloadCommand","ngc registry resource download-version nvidia/dle/jasper_pyt:20.10.0","helmFetchCommand","viewAllVersionsPath","/orgs/nvidia/dle/resources/jasper_pyt/-/version-history","accessCheckResult",{"_210":91},"hasRelatedCollections","context",{"_206":207,"_208":209},1787560659379,[],[],[199,200,201,202,203,204,205],{"_48":63,"_50":74,"_52":65,"_54":66,"_56":75,"_58":59},{"_48":63,"_50":72,"_52":147,"_54":-5,"_56":-5,"_58":59},{"_48":63,"_50":70,"_52":65,"_54":66,"_56":71,"_58":59},{"_48":63,"_50":68,"_52":65,"_54":66,"_56":69,"_58":59},{"_48":63,"_50":64,"_52":65,"_54":66,"_56":67,"_58":59},{"_48":60,"_50":61,"_52":53,"_54":55,"_56":62,"_58":59},{"_48":49,"_50":51,"_52":53,"_54":55,"_56":57,"_58":59},"orgName","nvidia","teamName","dle","allowed","versionId","description",["D",1663857364140],"totalSizeInBytes","malwareScanStatus","releaseNotes","We're constantly refining and improving our performance on AI and HPC workloads even on the same hardware with frequent updates to our software stack. For our latest performance data please refer to these pages for AI and HPC benchmarks.\n\n### Changelog\nFebruary 2021\n* Added DALI data-processing pipeline for on-the-fly data processing and augmentation on CPU or GPU\n* Revised training recipe: ~10% relative improvement in Word Error Rate (WER)\n* Updated Triton scripts for compatibility with Triton V2 API, updated Triton inference results\n* Refactored codebase\n* Updated performance results for the PyTorch 20.10-py3 NGC container\n\nJune 2020\n* Updated performance tables to include A100 results\n\nDecember 2019\n* Inference support for TRT 6 with dynamic shapes\n* Inference support for TensorRT Inference Server with acoustic model backends in ONNX, PyTorch JIT, TensorRT\n* Jupyter notebook for inference with TensorRT Inference Server\n\nNovember 2019\n* Google Colab notebook for inference with native TensorRT\n\nSeptember 2019\n* Inference support for TensorRT 6 with static shapes\n* Jupyter notebook for inference\n\nAugust 2019\n* Initial release\n\n### Known issues\nThere are no known issues in this release.","hasReleaseNotes","setup","The following section lists the requirements in order to start training and evaluating the Jasper model.\n\n### Requirements\nThis repository contains a `Dockerfile` which extends the PyTorch 20.10-py3 NGC container and encapsulates some dependencies. Aside from these dependencies, ensure you have the following components:\n\n* [NVIDIA Docker](https://github.com/NVIDIA/nvidia-docker)\n* [PyTorch 20.10-py3 NGC container](https://ngc.nvidia.com/catalog/containers/nvidia:pytorch)\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\nFurther required python packages are listed in `requirements.txt`, which are automatically installed with the Docker container built. To manually install them, run\n```bash\npip install -r requirements.txt\n```\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\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/dgx/user-guide/index.html#accessing_registry)\n* [Running PyTorch](https://docs.nvidia.com/deeplearning/dgx/pytorch-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/dgx/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 Jasper model on the Librispeech dataset. For details concerning training and inference, see [Advanced](#Advanced) section.\n\n1. Clone the repository.\n```bash\ngit clone https://github.com/NVIDIA/DeepLearningExamples\ncd DeepLearningExamples/PyTorch/SpeechRecognition/Jasper\n```\n2. Build the Jasper PyTorch container.\n\nRunning the following scripts will build and launch the container which contains all the required dependencies for data download and processing as well as training and inference of the model.\n\n```bash\nbash scripts/docker/build.sh\n```\n\n3. Start an interactive session in the NGC container to run data download/training/inference\n\n```bash\nbash scripts/docker/launch.sh \n```\nWithin the container, the contents of this repository will be copied to the `/workspace/jasper` directory. The `/datasets`, `/checkpoints`, `/results` directories are mounted as volumes\nand mapped to the corresponding directories ``, ``, `` on the host.\n\n4. Download and preprocess the dataset.\n\nNo GPU is required for data download and preprocessing. Therefore, if GPU usage is a limited resource, launch the container for this section on a CPU machine by following Steps 2 and 3.\n\nNote: Downloading and preprocessing the dataset requires 500GB of free disk space and can take several hours to complete.\n\nThis repository provides scripts to download, and extract the following datasets:\n\n* LibriSpeech [http://www.openslr.org/12](http://www.openslr.org/12)\n\nLibriSpeech contains 1000 hours of 16kHz read English speech derived from public domain audiobooks from LibriVox project and has been carefully segmented and aligned. For more information, see the [LIBRISPEECH: AN ASR CORPUS BASED ON PUBLIC DOMAIN AUDIO BOOKS](http://www.danielpovey.com/files/2015_icassp_librispeech.pdf) paper.\n\nInside the container, download and extract the datasets into the required format for later training and inference:\n```bash\nbash scripts/download_librispeech.sh\n```\nOnce the data download is complete, the following folders should exist:\n```bash\ndatasets/LibriSpeech/\n|-- dev-clean\n|-- dev-other\n|-- test-clean\n|-- test-other\n|-- train-clean-100\n|-- train-clean-360\n|-- train-other-500\n```\n\nSince `/datasets/` is mounted to `` on the host (see Step 3), once the dataset is downloaded it will be accessible from outside of the container at `/LibriSpeech`.\n\nNext, convert the data into WAV files:\n```bash\nbash scripts/preprocess_librispeech.sh\n```\nOnce the data is converted, the following additional files and folders should exist:\n```bash\ndatasets/LibriSpeech/\n|-- dev-clean-wav\n|-- dev-other-wav\n|-- librispeech-train-clean-100-wav.json\n|-- librispeech-train-clean-360-wav.json\n|-- librispeech-train-other-500-wav.json\n|-- librispeech-dev-clean-wav.json\n|-- librispeech-dev-other-wav.json\n|-- librispeech-test-clean-wav.json\n|-- librispeech-test-other-wav.json\n|-- test-clean-wav\n|-- test-other-wav\n|-- train-clean-100-wav\n|-- train-clean-360-wav\n|-- train-other-500-wav\n```\n\nThe DALI data pre-processing pipeline, which is enabled by default, performs speed perturbation on-line during training.\nWithout DALI, on-line speed perturbation might slow down the training.\nIf you wish to disable DALI, speed perturbation can be computed off-line with:\n```bash\nSPEEDS=\"0.9 1.1\" bash scripts/preprocess_librispeech.sh\n```\n\n5. Start training.\n\nInside the container, use the following script to start training.\nMake sure the downloaded and preprocessed dataset is located at `/LibriSpeech` on the host (see Step 3), which corresponds to `/datasets/LibriSpeech` inside the container.\n\n```bash\n[OPTION1=value1 OPTION2=value2 ...] bash scripts/train.sh\n```\nBy default automatic precision is disabled, batch size is 64 over two gradient accumulation steps, and the recipe is run on a total of 8 GPUs. The hyperparameters are tuned for a GPU with at least 32GB of memory and will require adjustment for 16GB GPUs (e.g., by lowering batch size and using more gradient accumulation steps).\n\nOptions are being passed as environment variables. More details on available options can be found in [Parameters](#parameters) and [Training process](#training-process).\n\n6. Start validation/evaluation.\n\nInside the container, use the following script to run evaluation.\n Make sure the downloaded and preprocessed dataset is located at `/LibriSpeech` on the host (see Step 3), which corresponds to `/datasets/LibriSpeech` inside the container.\n```bash\n[OPTION1=value1 OPTION2=value2 ...] bash scripts/evaluation.sh [OPTIONS]\n```\nBy default, this will use full precision, a batch size of 64 and run on a single GPU.\n\nOptions are being passed as environment variables. More details on available options can be found in [Parameters](#parameters) and [Evaluation process](#evaluation-process).\n\n7. Start inference/predictions.\n\nInside the container, use the following script to run inference.\n Make sure the downloaded and preprocessed dataset is located at `/LibriSpeech` on the host (see Step 3), which corresponds to `/datasets/LibriSpeech` inside the container.\nA pretrained model checkpoint can be downloaded from [NGC model repository](https://ngc.nvidia.com/catalog/models/nvidia:jasperpyt_fp16).\n\n```bash\n[OPTION1=value1 OPTION2=value2 ...] bash scripts/inference.sh\n```\nBy default this will use single precision, a batch size of 64 and run on a single GPU.\n\nOptions are being passed as environment variables. More details on available options can be found in [Parameters](#parameters) and [Inference process](#inference-process).","hasQuickStartGuide","advanced","The following sections provide greater details of the dataset, running training and inference, and getting training and inference results.\n\n### Scripts and sample code\nIn the `root` directory, the most important files are:\n```\njasper\n|-- common # data pre-processing, logging, etc.\n|-- configs # model configurations\n|-- Dockerfile # container with the basic set of dependencies to run Jasper\n|-- inference.py # entry point for inference\n|-- jasper # model-specific code\n|-- notebooks # jupyter notebooks and example audio files\n|-- scripts # one-click scripts required for running various supported functionalities\n│ |-- docker # contains the scripts for building and launching the container\n│ |-- download_librispeech.sh # downloads LibriSpeech dataset\n│ |-- evaluation.sh # runs evaluation using the `inference.py` script\n│ |-- inference_benchmark.sh # runs the inference benchmark using the `inference_benchmark.py` script\n│ |-- inference.sh # runs inference using the `inference.py` script\n│ |-- preprocess_librispeech.sh # preprocess LibriSpeech raw data files for training and inference\n│ |-- train_benchmark.sh # runs the training performance benchmark using the `train.py` script\n│ |-- train.sh # runs training using the `train.py` script\n|-- train.py # entry point for training\n|-- triton # example of inference using Triton Inference Server\n|-- utils # data downloading and common routines\n```\n\n### Parameters\n\nParameters could be set as env variables, or passed as positional arguments.\n\nThe complete list of available parameters for `scripts/train.sh` script contains:\n```bash\nDATA_DIR: directory of dataset. (default: '/datasets/LibriSpeech')\nMODEL_CONFIG: relative path to model configuration. (default: 'configs/jasper10x5dr_speedp-online_speca.yaml')\nOUTPUT_DIR: directory for results, logs, and created checkpoints. (default: '/results')\nCHECKPOINT: a specific model checkpoint to continue training from. To resume training from the last checkpoint, see the RESUME option.\nRESUME: resume training from the last checkpoint found in OUTPUT_DIR, or from scratch if there are no checkpoints (default: true)\nCUDNN_BENCHMARK: boolean that indicates whether to enable cudnn benchmark mode for using more optimized kernels. (default: true)\nNUM_GPUS: number of GPUs to use. (default: 8)\nAMP: if set to `true`, enables automatic mixed precision (default: false)\nBATCH_SIZE: effective data batch size. The real batch size per GPU might be lower, if gradient accumulation is enabled (default: 64)\nGRAD_ACCUMULATION_STEPS: number of gradient accumulation steps until optimizer updates weights. (default: 2)\nLEARNING_RATE: initial learning rate. (default: 0.01)\nMIN_LEARNING_RATE: minimum learning rate, despite LR scheduling (default: 1e-5)\nLR_POLICY: how to decay LR (default: exponential)\nLR_EXP_GAMMA: decay factor for the exponential LR schedule (default: 0.981)\nEMA: decay factor for exponential averages of checkpoints (default: 0.999)\nSEED: seed for random number generator and used for ensuring reproducibility. (default: 0)\nEPOCHS: number of training epochs. (default: 440)\nWARMUP_EPOCHS: number of initial epoch of linearly increasing LR. (default: 2)\nHOLD_EPOCHS: number of epochs to hold maximum LR after warmup. (default: 140)\nSAVE_FREQUENCY: number of epochs between saving the model to disk. (default: 10)\nEPOCHS_THIS_JOB: run training for this number of epochs. Does not affect LR schedule like the EPOCHS parameter. (default: 0)\nDALI_DEVICE: device to run the DALI pipeline on for calculation of filterbanks. Valid choices: cpu, gpu, none. (default: gpu)\nPAD_TO_MAX_DURATION: pad all sequences with zeros to maximum length. (default: false)\nEVAL_FREQUENCY: number of steps between evaluations on the validation set. (default: 544)\nPREDICTION_FREQUENCY: the number of steps between writing a sample prediction to stdout. (default: 544)\nTRAIN_MANIFESTS: lists of .json training set files\nVAL_MANIFESTS: lists of .json validation set files\n\n```\n\nThe complete list of available parameters for `scripts/inference.sh` script contains:\n```bash\nDATA_DIR: directory of dataset. (default: '/datasets/LibriSpeech')\nMODEL_CONFIG: model configuration. (default: 'configs/jasper10x5dr_speedp-online_speca.yaml')\nOUTPUT_DIR: directory for results and logs. (default: '/results')\nCHECKPOINT: model checkpoint path. (required)\nDATASET: name of the LibriSpeech subset to use. (default: 'dev-clean')\nLOG_FILE: path to the DLLogger .json logfile. (default: '')\nCUDNN_BENCHMARK: enable cudnn benchmark mode for using more optimized kernels. (default: false)\nMAX_DURATION: filter out recordings shorter then MAX_DURATION seconds. (default: \"\")\nPAD_TO_MAX_DURATION: pad all sequences with zeros to maximum length. (default: false)\nPAD_LEADING: pad every batch with leading zeros to counteract conv shifts of the field of view. (default: 16)\nNUM_GPUS: number of GPUs to use. Note that with > 1 GPUs WER results might be inaccurate due to the batching policy. (default: 1)\nNUM_STEPS: number of batches to evaluate, loop the dataset if necessary. (default: 0)\nNUM_WARMUP_STEPS: number of initial steps before measuring performance. (default: 0)\nAMP: enable FP16 inference with AMP. (default: false)\nBATCH_SIZE: data batch size. (default: 64)\nEMA: Attempt to load exponentially averaged weights from a checkpoint. (default: true)\nSEED: seed for random number generator and used for ensuring reproducibility. (default: 0)\nDALI_DEVICE: device to run the DALI pipeline on for calculation of filterbanks. Valid choices: cpu, gpu, none. (default: gpu)\nCPU: run inference on CPU. (default: false)\nLOGITS_FILE: dump logit matrices to a file. (default: \"\")\nPREDICTION_FILE: save predictions to a file. (default: \"${OUTPUT_DIR}/${DATASET}.predictions\")\n```\n\nThe complete list of available parameters for `scripts/evaluation.sh` is the same as for `scripts/inference.sh` except for the few default changes.\n```bash\nPREDICTION_FILE: (default: \"\")\nDATASET: (default: \"test-other\")\n```\n\nThe `scripts/inference_benchmark.sh` script pads all input to a fixed duration and computes the mean, 90%, 95%, 99% percentile of latency for the specified number of inference steps. Latency is measured in milliseconds per batch. The `scripts/inference_benchmark.sh` measures latency for a single GPU and loops over a number of batch sizes and durations. It extends `scripts/inference.sh`, and changes the defaults with:\n```bash\nBATCH_SIZE_SEQ: batch sizes to measure on. (default: \"1 2 4 8 16\")\nMAX_DURATION_SEQ: input durations (in seconds) to measure on (default: \"2 7 16.7\")\nCUDNN_BENCHMARK: (default: true)\nPAD_TO_MAX_DURATION: (default: true)\nPAD_LEADING: (default: 0)\nNUM_WARMUP_STEPS: (default: 10)\nNUM_STEPS: (default: 500)\nDALI_DEVICE: (default: cpu)\n```\n\nThe `scripts/train_benchmark.sh` script pads all input to the same length according to the input argument `MAX_DURATION` and measures average training latency and throughput performance. Latency is measured in seconds per batch, throughput in sequences per second.\nTraining performance is measured with on-line speed perturbation and cuDNN benchmark mode enabled.\nThe script `scripts/train_benchmark.sh` loops over a number of batch sizes and GPU counts.\nIt extends `scripts/train.sh`, and the complete list of available parameters for `scripts/train_benchmark.sh` script contains:\n```bash\nBATCH_SIZE_SEQ: batch sizes to measure on. (default: \"1 2 4 8 16\")\nNUM_GPUS_SEQ: number of GPUs to run the training on. (default: \"1 4 8\")\nMODEL_CONFIG: (default: \"configs/jasper10x5dr_speedp-online_train-benchmark.yaml\")\nTRAIN_MANIFESTS: (default: \"$DATA_DIR/librispeech-train-clean-100-wav.json\")\nRESUME: (default: false)\nEPOCHS_THIS_JOB: (default: 2)\nEPOCHS: (default: 100000)\nSAVE_FREQUENCY: (default: 100000)\nEVAL_FREQUENCY: (default: 100000)\nGRAD_ACCUMULATION_STEPS: (default: 1)\nPAD_TO_MAX_DURATION: (default: true)\nEMA: (default: 0)\n```\n\n### Command-line options\nTo see the full list of available options and their descriptions, use the `-h` or `--help` command-line option with the Python file, for example:\n\n```bash\npython train.py --help\npython inference.py --help\n```\n\n### Getting the data\nThe Jasper model was trained on the LibriSpeech dataset. We use the concatenation of `train-clean-100`, `train-clean-360` and `train-other-500` for training and `dev-clean` for validation.\n\nThis repository contains the `scripts/download_librispeech.sh` and `scripts/preprocess_librispeech.sh` scripts which will automatically download and preprocess the training, test and development datasets. By default, data will be downloaded to the `/datasets/LibriSpeech` directory, a minimum of 250GB free space is required for download and preprocessing, the final preprocessed dataset is approximately 100GB. With offline speed perturbation, the dataset will be about 3x larger.\n\n#### Dataset guidelines\nThe `scripts/preprocess_librispeech.sh` script converts the input audio files to WAV format with a sample rate of 16kHz, target transcripts are stripped from whitespace characters, then lower-cased. For `train-clean-100`, `train-clean-360` and `train-other-500`. It can optionally create speed perturbed versions with rates of 0.9 and 1.1 for data augmentation. In the current version, those augmentations are applied on-line with the DALI pipeline without any impact on training time.\n\nAfter preprocessing, the script creates JSON files with output file paths, sample rate, target transcript and other metadata. These JSON files are used by the training script to identify training and validation datasets.\n\nThe Jasper model was tuned on audio signals with a sample rate of 16kHz, if you wish to use a different sampling rate then some hyperparameters might need to be changed - specifically window size and step size.\n\n### Training process\n\nThe training is performed using `train.py` script along with parameters defined in `scripts/train.sh`\nThe `scripts/train.sh` script runs a job on a single node that trains the Jasper model from scratch using LibriSpeech as training data. To make training more efficient, we discard audio samples longer than 16.7 seconds from the training dataset, the total number of these samples is less than 1%. Such filtering does not degrade accuracy, but it allows us to decrease the number of time steps in a batch, which requires less GPU memory and increases training speed.\nApart from the default arguments as listed in the [Parameters](#parameters) section, by default the training script:\n\n* Runs on 8 GPUs with at least 32GB of memory and training/evaluation batch size 64, split over two gradient accumulation steps\n* Uses TF32 precision (A100 GPU) or FP32 (other GPUs)\n* Trains on the concatenation of all 3 LibriSpeech training datasets and evaluates on the LibriSpeech dev-clean dataset\n* Maintains an exponential moving average of parameters for evaluation\n* Has cudnn benchmark enabled\n* Runs for 440 epochs\n* Uses an initial learning rate of 0.01 and an exponential learning rate decay\n* Saves a checkpoint every 10 epochs\n* Automatically removes old checkpoints and preserves milestone checkpoints\n* Runs evaluation on the development dataset every 544 iterations and at the end of training\n* Maintains a separate checkpoint with the lowest WER on development set\n* Prints out training progress every iteration to stdout\n* Creates a DLLogger logfile and a Tensorboard log\n* Calculates speed perturbation on-line during training\n* Uses SpecAugment in data pre-processing\n* Filters out audio samples longer than 16.7 seconds\n* Pads each batch so its length would be divisible by 16\n* Uses masked convolutions and dense residuals as described in the paper\n* Uses weight decay of 0.001\n* Uses [Novograd](https://arxiv.org/pdf/1905.11286.pdf) as optimizer with betas=(0.95, 0)\n\nEnabling AMP permits batch size 64 with one gradient accumulation step. In the current setup it will improve upon the greedy WER [Results](#results) of the Jasper paper on a DGX-1 with 32GB V100 GPUs.\n\n### Inference process\nInference is performed using the `inference.py` script along with parameters defined in `scripts/inference.sh`.\nThe `scripts/inference.sh` script runs the job on a single GPU, taking a pre-trained Jasper model checkpoint and running it on the specified dataset.\nApart from the default arguments as listed in the [Parameters](#parameters) section by default the inference script:\n\n* Evaluates on the LibriSpeech dev-clean dataset\n* Uses a batch size of 64\n* Runs for 1 epoch and prints out the final word error rate\n* Creates a log file with progress and results which will be stored in the results folder\n* Pads each batch so its length would be divisible by 16\n* Does not use data augmentation\n* Does greedy decoding and saves the transcription in the results folder\n* Has the option to save the model output tensors for more complex decoding, for example, beam search\n* Has cudnn benchmark disabled\n\n### Evaluation process\nEvaluation is performed using the `inference.py` script along with parameters defined in `scripts/evaluation.sh`.\nThe setup is similar to `scripts/inference.sh`, with two differences:\n\n* Evaluates the LibriSpeech test-other dataset\n* Model outputs are not saved\n\n### Deploying Jasper using Triton Inference Server\nThe NVIDIA Triton Inference Server provides a datacenter and cloud inferencing solution optimized for NVIDIA GPUs. The server provides an inference service via an HTTP or gRPC endpoint, allowing remote clients to request inferencing for any number of GPU or CPU models being managed by the server.\nMore information on how to perform inference using Triton Inference Server with different model backends can be found in the subfolder [./https://raw.githubusercontent.com/NVIDIA/DeepLearningExamples/master/PyTorch/SpeechRecognition/Jasper//triton/README.md](https://raw.githubusercontent.com/NVIDIA/DeepLearningExamples/master/PyTorch/SpeechRecognition/Jasper//triton/README.md)","hasAdvanced","performance","The performance measurements in this document were conducted at the time of publication and may not reflect the performance achieved from NVIDIA's latest software release. For the most up-to-date performance measurements, go to [NVIDIA Data Center Deep Learning Product Performance](https://developer.nvidia.com/deep-learning-performance-training-inference).\n\n### Benchmarking\nThe following section shows how to run benchmarks measuring the model performance in training and inference modes.\n\n#### Training performance benchmark\nTo benchmark the training performance in a specific setting on the `train-clean-100` subset of LibriSpeech, run:\n\n```bash\nBATCH_SIZE_SEQ= NUM_GPUS_SEQ= bash scripts/train_benchmark.sh\n```\n\nBy default, this script runs 2 epochs on the configuration `configs/jasper10x5dr_speedp-online_train-benchmark.yaml`,\nwhich applies gentle speed perturbation that does not change the length of the output, enabling immediate stabilization of training step times in the cuDNN benchmark mode. The script runs benchmarks on batch sizes 32 on 1, 4, and 8 GPUs, and requires a 8x 32GB GPU machine.\n\n#### Inference performance benchmark\nTo benchmark the inference performance on a specific batch size and audio length, run:\n\n```bash\nBATCH_SIZE_SEQ= MAX_DURATION_SEQ= bash scripts/inference_benchmark.sh\n```\nBy default, the script runs on a single GPU and evaluates on the dataset limited to utterances shorter than MAX_DURATION. It uses the model configuration `configs/jasper10x5dr_speedp-online_speca.yaml`.\n\n### Results\nThe following sections provide details on how we achieved our performance and accuracy in training and inference.\nAll results are trained on 960 hours of LibriSpeech with a maximum audio length of 16.7s. The training is evaluated\non LibriSpeech dev-clean, dev-other, test-clean, test-other. Checkpoints for evaluation are being chosen based on their\nword error rate on dev-clean.\n\n#### Training accuracy results\n\n##### Training accuracy: NVIDIA DGX A100 (8x A100 80GB)\nOur results were obtained by running the `scripts/train.sh` training script in the PyTorch 20.10-py3 NGC container with NVIDIA DGX A100 with (8x A100 80GB) GPUs.\nThe following table reports the word error rate (WER) of the acoustic model with greedy decoding on all LibriSpeech dev and test datasets for mixed precision training.\n\n| Number of GPUs | Batch size per GPU | Precision | dev-clean WER | dev-other WER | test-clean WER | test-other WER | Time to train |\n|-----|-----|-------|-------|-------|------|-------|------|\n| 8 | 64 | mixed | 3.20 | 9.78 | 3.41 | 9.71 | 70 h |\n\n##### Training accuracy: NVIDIA DGX-1 (8x V100 32GB)\nOur results were obtained by running the `scripts/train.sh` training script in the PyTorch 20.10-py3 NGC container with NVIDIA DGX-1 with (8x V100 32GB) GPUs.\nThe following table reports the word error rate (WER) of the acoustic model with greedy decoding on all LibriSpeech dev and test datasets for mixed precision training.\n\n| Number of GPUs | Batch size per GPU | Precision | dev-clean WER | dev-other WER | test-clean WER | test-other WER | Time to train |\n|-----|-----|-------|-------|-------|------|-------|-------|\n| 8 | 64 | mixed | 3.26 | 10.00 | 3.54 | 9.80 | 130 h |\n\nWe show the best of 5 runs (mixed precision) and 2 runs (FP32) chosen based on dev-clean WER. For FP32, two gradient accumulation steps have been used.\n\n##### Training stability test\nThe following table compares greedy decoding word error rates across 8 different training runs with different seeds for mixed precision training.\n\n| DGX A100 80GB, FP16, 8x GPU | Seed #1 | Seed #2 | Seed #3 | Seed #4 | Seed #5 | Seed #6 | Seed #7 | Seed #8 | Mean | Std |\n|-----------:|----------:|----------:|----------:|----------:|----------:|----------:|----------:|----------:|-------:|------:|\n| dev-clean | 3.46 | 3.55 | 3.45 | 3.44 | 3.25 | 3.34 | 3.20 | 3.40 | 3.39 | 0.11 |\n| dev-other | 10.30 | 10.77 | 10.36 | 10.26 | 9.99 | 10.18 | 9.78 | 10.32 | 10.25 | 0.27 |\n| test-clean | 3.84 | 3.81 | 3.66 | 3.64 | 3.58 | 3.55 | 3.41 | 3.73 | 3.65 | 0.13 |\n| test-other | 10.61 | 10.52 | 10.49 | 10.47 | 9.89 | 10.09 | 9.71 | 10.26 | 10.26 | 0.31 |\n\n| DGX-1 32GB, FP16, 8x GPU | Seed #1 | Seed #2 | Seed #3 | Seed #4 | Seed #5 | Seed #6 | Seed #7 | Seed #8 | Mean | Std |\n|-----------:|----------:|----------:|----------:|----------:|----------:|----------:|----------:|----------:|-------:|------:|\n| dev-clean | 3.31 | 3.31 | 3.26 | 3.44 | 3.40 | 3.35 | 3.36 | 3.28 | 3.34 | 0.06 |\n| dev-other | 10.02 | 10.01 | 10.00 | 10.06 | 10.05 | 10.03 | 10.10 | 10.04 | 10.04 | 0.03 |\n| test-clean | 3.49 | 3.50 | 3.54 | 3.61 | 3.57 | 3.58 | 3.48 | 3.51 | 3.54 | 0.04 |\n| test-other | 10.11 | 10.14 | 9.80 | 10.09 | 10.17 | 9.99 | 9.86 | 10.00 | 10.02 | 0.13 |\n\n#### Training performance results\nOur results were obtained by running the `scripts/train.sh` training script in the PyTorch 20.10-py3 NGC container. Performance (in sequences per second) is the steady-state throughput.\n\n##### Training performance: NVIDIA DGX A100 (8x A100 80GB)\n\n| Batch size / GPU | GPUs | Throughput - TF32 | Throughput - mixed precision | Throughput speedup (TF32 to mixed precision) | Weak scaling - TF32 | Weak scaling - mixed precision |\n|----:|----:|-------:|-------:|-----:|-----:|-----:|\n| 32 | 1 | 42.18 | 64.32 | 1.52 | 1.00 | 1.00 |\n| 32 | 4 | 157.49 | 239.23 | 1.52 | 3.73 | 3.72 |\n| 32 | 8 | 310.10 | 470.09 | 1.52 | 7.35 | 7.31 |\n| 64 | 1 | 49.64 | 75.59 | 1.52 | 1.00 | 1.00 |\n| 64 | 4 | 192.66 | 289.16 | 1.50 | 3.88 | 3.83 |\n| 64 | 8 | 371.41 | 547.91 | 1.48 | 7.48 | 7.25 |\n\nNote: Mixed precision permits higher batch sizes during training. We report the maximum batch sizes (as powers of 2), which are allowed without gradient accumulation.\n\nTo achieve these same results, follow the [Quick Start Guide](#quick-start-guide) outlined above.\n\n##### Training performance: NVIDIA DGX-1 (8x V100 16GB)\n\n| Batch size / GPU | GPUs | Throughput - FP32 | Throughput - mixed precision | Throughput speedup (FP32 to mixed precision) | Weak scaling - FP32 | Weak scaling - mixed precision |\n|----:|----:|------:|-------:|-----:|-----:|-----:|\n| 16 | 1 | 10.71 | 27.87 | 2.60 | 1.00 | 1.00 |\n| 16 | 4 | 40.28 | 99.80 | 2.48 | 3.76 | 3.58 |\n| 16 | 8 | 78.23 | 193.89 | 2.48 | 7.30 | 6.96 |\n\nNote: Mixed precision permits higher batch sizes during training. We report the maximum batch sizes (as powers of 2), which are allowed without gradient accumulation.\n\nTo achieve these same results, follow the [Quick Start Guide](#quick-start-guide) outlined above.\n\n##### Training performance: NVIDIA DGX-1 (8x V100 32GB)\n\n| Batch size / GPU | GPUs | Throughput - FP32 | Throughput - mixed precision | Throughput speedup (FP32 to mixed precision) | Weak scaling - FP32 | Weak scaling - mixed precision |\n|----:|----:|------:|-------:|-----:|-----:|-----:|\n| 32 | 1 | 12.22 | 34.08 | 2.79 | 1.00 | 1.00 |\n| 32 | 4 | 46.97 | 128.39 | 2.73 | 3.84 | 3.77 |\n| 32 | 8 | 92.44 | 249.00 | 2.69 | 7.57 | 7.31 |\n| 64 | 1 | N/A | 39.30 | N/A | N/A | 1.00 |\n| 64 | 4 | N/A | 150.18 | N/A | N/A | 3.82 |\n| 64 | 8 | N/A | 282.68 | N/A | N/A | 7.19 |\n\nNote: Mixed precision permits higher batch sizes during training. We report the maximum batch sizes (as powers of 2), which are allowed without gradient accumulation.\n\nTo achieve these same results, follow the [Quick Start Guide](#quick-start-guide) outlined above.\n\n##### Training performance: NVIDIA DGX-2 (16x V100 32GB)\n\n| Batch size / GPU | GPUs | Throughput - FP32 | Throughput - mixed precision | Throughput speedup (FP32 to mixed precision) | Weak scaling - FP32 | Weak scaling - mixed precision |\n|----:|----:|-------:|-------:|-----:|------:|------:|\n| 32 | 1 | 13.46 | 38.94 | 2.89 | 1.00 | 1.00 |\n| 32 | 4 | 51.38 | 143.44 | 2.79 | 3.82 | 3.68 |\n| 32 | 8 | 100.54 | 280.48 | 2.79 | 7.47 | 7.20 |\n| 32 | 16 | 188.14 | 515.90 | 2.74 | 13.98 | 13.25 |\n| 64 | 1 | N/A | 43.86 | N/A | N/A | 1.00 |\n| 64 | 4 | N/A | 165.27 | N/A | N/A | 3.77 |\n| 64 | 8 | N/A | 318.10 | N/A | N/A | 7.25 |\n| 64 | 16 | N/A | 567.47 | N/A | N/A | 12.94 |\n\nNote: Mixed precision permits higher batch sizes during training. We report the maximum batch sizes (as powers of 2), which are allowed without gradient accumulation.\n\nTo achieve these same results, follow the [Quick Start Guide](#quick-start-guide) outlined above.\n\n#### Inference performance results\nOur results were obtained by running the `scripts/inference_benchmark.sh` script in the PyTorch 20.10-py3 NGC container on NVIDIA DGX A100, DGX-1, DGX-2 and T4 on a single GPU. Performance numbers (latency in milliseconds per batch) were averaged over 500 iterations.\n\n##### Inference performance: NVIDIA DGX A100 (1x A100 80GB)\n\n| | | FP16 Latency (ms) Percentiles | | | | TF32 Latency (ms) Percentiles | | | | FP16/TF32 speed up |\n|-----:|---------------:|------:|------:|------:|------:|------:|------:|-------:|------:|------:|\n| BS | Duration (s) | 90% | 95% | 99% | Avg | 90% | 95% | 99% | Avg | Avg |\n| 1 | 2.0 | 32.40 | 32.50 | 32.82 | 32.30 | 33.30 | 33.64 | 34.65 | 33.25 | 1.03 |\n| 2 | 2.0 | 32.90 | 33.51 | 34.35 | 32.69 | 34.48 | 34.65 | 35.66 | 34.27 | 1.05 |\n| 4 | 2.0 | 32.85 | 33.01 | 33.89 | 32.60 | 34.09 | 34.46 | 35.22 | 34.00 | 1.04 |\n| 8 | 2.0 | 35.51 | 35.89 | 37.10 | 35.33 | 34.86 | 35.36 | 36.08 | 34.45 | 0.98 |\n| 16 | 2.0 | 36.00 | 36.57 | 37.40 | 35.77 | 43.83 | 44.12 | 44.77 | 43.39 | 1.21 |\n| 1 | 7.0 | 33.50 | 33.99 | 34.91 | 33.03 | 33.83 | 34.25 | 34.95 | 33.70 | 1.02 |\n| 2 | 7.0 | 34.43 | 34.89 | 35.72 | 34.22 | 34.41 | 34.73 | 35.69 | 34.28 | 1.00 |\n| 4 | 7.0 | 34.30 | 34.59 | 35.43 | 34.07 | 37.95 | 38.18 | 38.87 | 37.55 | 1.10 |\n| 8 | 7.0 | 35.98 | 36.28 | 37.11 | 35.28 | 44.64 | 44.79 | 45.37 | 44.29 | 1.26 |\n| 16 | 7.0 | 39.86 | 40.08 | 41.16 | 39.33 | 55.17 | 55.46 | 57.24 | 54.56 | 1.39 |\n| 1 | 16.7 | 35.20 | 35.80 | 38.71 | 34.36 | 35.36 | 35.76 | 36.55 | 34.64 | 1.01 |\n| 2 | 16.7 | 35.40 | 35.81 | 36.50 | 34.76 | 36.34 | 36.53 | 37.40 | 35.87 | 1.03 |\n| 4 | 16.7 | 36.01 | 36.38 | 37.37 | 35.57 | 44.69 | 45.09 | 45.88 | 43.92 | 1.23 |\n| 8 | 16.7 | 41.48 | 41.78 | 44.22 | 40.69 | 58.57 | 58.74 | 59.62 | 58.11 | 1.43 |\n| 16 | 16.7 | 61.37 | 61.93 | 66.32 | 60.92 | 97.33 | 97.71 | 100.04 | 96.56 | 1.59 |\n\nTo achieve these same results, follow the [Quick Start Guide](#quick-start-guide) outlined above.\n\n##### Inference performance: NVIDIA DGX-1 (1x V100 16GB)\n\n| | | FP16 Latency (ms) Percentiles | | | | FP32 Latency (ms) Percentiles | | | | FP16/FP32 speed up |\n|-----:|---------------:|-------:|-------:|-------:|-------:|-------:|-------:|-------:|-------:|------:|\n| BS | Duration (s) | 90% | 95% | 99% | Avg | 90% | 95% | 99% | Avg | Avg |\n| 1 | 2.0 | 45.42 | 45.62 | 49.54 | 45.02 | 48.83 | 48.99 | 51.66 | 48.44 | 1.08 |\n| 2 | 2.0 | 50.31 | 50.53 | 53.66 | 49.10 | 49.87 | 50.04 | 52.99 | 49.41 | 1.01 |\n| 4 | 2.0 | 49.17 | 49.48 | 52.13 | 48.73 | 52.92 | 53.21 | 55.28 | 52.31 | 1.07 |\n| 8 | 2.0 | 51.20 | 51.40 | 52.32 | 49.01 | 73.02 | 73.30 | 75.00 | 71.99 | 1.47 |\n| 16 | 2.0 | 51.75 | 52.24 | 56.36 | 51.27 | 83.99 | 84.57 | 86.69 | 83.24 | 1.62 |\n| 1 | 7.0 | 48.13 | 48.53 | 50.95 | 46.78 | 48.52 | 48.75 | 50.89 | 48.01 | 1.03 |\n| 2 | 7.0 | 49.52 | 50.10 | 52.35 | 48.00 | 65.27 | 65.41 | 66.59 | 64.79 | 1.35 |\n| 4 | 7.0 | 51.75 | 52.01 | 54.39 | 50.38 | 93.75 | 94.77 | 97.04 | 92.27 | 1.83 |\n| 8 | 7.0 | 54.80 | 56.27 | 66.23 | 52.95 | 130.65 | 131.09 | 132.91 | 129.82 | 2.45 |\n| 16 | 7.0 | 73.02 | 73.42 | 75.83 | 71.96 | 157.53 | 158.20 | 160.73 | 155.51 | 2.16 |\n| 1 | 16.7 | 48.10 | 48.52 | 52.71 | 47.20 | 73.34 | 73.56 | 74.19 | 72.69 | 1.54 |\n| 2 | 16.7 | 64.21 | 64.52 | 65.56 | 56.06 | 129.48 | 129.97 | 131.78 | 126.36 | 2.25 |\n| 4 | 16.7 | 60.38 | 61.03 | 63.18 | 58.87 | 183.33 | 183.85 | 185.53 | 181.90 | 3.09 |\n| 8 | 16.7 | 85.88 | 86.34 | 87.70 | 84.46 | 227.42 | 228.21 | 229.63 | 225.71 | 2.67 |\n| 16 | 16.7 | 135.62 | 136.40 | 137.69 | 131.58 | 276.90 | 277.59 | 281.16 | 275.08 | 2.09 |\n\nTo achieve these same results, follow the [Quick Start Guide](#quick-start-guide) outlined above.\n\n##### Inference performance: NVIDIA DGX-1 (1x V100 32GB)\n\n| | | FP16 Latency (ms) Percentiles | | | | FP32 Latency (ms) Percentiles | | | | FP16/FP32 speed up |\n|-----:|---------------:|-------:|-------:|-------:|-------:|-------:|-------:|-------:|-------:|------:|\n| BS | Duration (s) | 90% | 95% | 99% | Avg | 90% | 95% | 99% | Avg | Avg |\n| 1 | 2.0 | 52.74 | 53.01 | 54.40 | 51.47 | 55.97 | 56.22 | 57.93 | 54.93 | 1.07 |\n| 2 | 2.0 | 51.77 | 52.15 | 54.69 | 50.98 | 56.58 | 56.87 | 58.88 | 55.35 | 1.09 |\n| 4 | 2.0 | 51.41 | 51.76 | 53.47 | 50.55 | 61.56 | 61.87 | 63.81 | 60.74 | 1.20 |\n| 8 | 2.0 | 51.83 | 52.15 | 54.08 | 50.85 | 80.20 | 80.69 | 81.67 | 77.69 | 1.53 |\n| 16 | 2.0 | 70.48 | 70.96 | 72.11 | 62.98 | 93.00 | 93.44 | 94.17 | 89.05 | 1.41 |\n| 1 | 7.0 | 49.77 | 50.21 | 51.88 | 48.73 | 52.74 | 52.99 | 54.54 | 51.67 | 1.06 |\n| 2 | 7.0 | 51.12 | 51.47 | 52.84 | 49.98 | 65.33 | 65.63 | 67.07 | 64.64 | 1.29 |\n| 4 | 7.0 | 53.13 | 53.56 | 55.68 | 52.15 | 93.54 | 93.85 | 94.72 | 92.76 | 1.78 |\n| 8 | 7.0 | 57.67 | 58.07 | 59.89 | 56.41 | 133.93 | 134.18 | 134.88 | 133.15 | 2.36 |\n| 16 | 7.0 | 76.09 | 76.48 | 79.13 | 75.27 | 162.35 | 162.77 | 164.63 | 161.30 | 2.14 |\n| 1 | 16.7 | 54.78 | 55.29 | 56.83 | 52.51 | 75.37 | 76.27 | 78.05 | 74.32 | 1.42 |\n| 2 | 16.7 | 56.80 | 57.20 | 59.01 | 55.49 | 130.60 | 131.36 | 132.93 | 128.55 | 2.32 |\n| 4 | 16.7 | 64.19 | 64.84 | 66.47 | 62.87 | 188.09 | 188.76 | 190.07 | 185.76 | 2.95 |\n| 8 | 16.7 | 87.46 | 87.86 | 89.99 | 86.47 | 232.33 | 232.89 | 234.43 | 230.44 | 2.67 |\n| 16 | 16.7 | 136.02 | 136.52 | 139.44 | 134.78 | 283.87 | 284.59 | 286.70 | 282.01 | 2.09 |\n\nTo achieve these same results, follow the [Quick Start Guide](#quick-start-guide) outlined above.\n\n##### Inference performance: NVIDIA DGX-2 (1x V100 32GB)\n\n| | | FP16 Latency (ms) Percentiles | | | | FP32 Latency (ms) Percentiles | | | | FP16/FP32 speed up |\n|-----:|---------------:|-------:|-------:|-------:|-------:|-------:|-------:|-------:|-------:|------:|\n| BS | Duration (s) | 90% | 95% | 99% | Avg | 90% | 95% | 99% | Avg | Avg |\n| 1 | 2.0 | 35.88 | 36.12 | 39.80 | 35.20 | 42.95 | 43.67 | 46.65 | 42.23 | 1.20 |\n| 2 | 2.0 | 36.36 | 36.57 | 40.97 | 35.60 | 41.83 | 42.21 | 45.60 | 40.97 | 1.15 |\n| 4 | 2.0 | 36.69 | 36.89 | 41.25 | 36.05 | 48.35 | 48.52 | 52.35 | 47.80 | 1.33 |\n| 8 | 2.0 | 37.49 | 37.70 | 41.37 | 36.88 | 65.41 | 65.64 | 66.50 | 64.96 | 1.76 |\n| 16 | 2.0 | 41.35 | 41.79 | 45.58 | 40.91 | 77.22 | 77.51 | 79.48 | 76.54 | 1.87 |\n| 1 | 7.0 | 36.07 | 36.55 | 40.31 | 35.62 | 39.52 | 39.84 | 43.07 | 38.93 | 1.09 |\n| 2 | 7.0 | 37.42 | 37.66 | 41.36 | 36.79 | 55.94 | 56.19 | 58.33 | 55.60 | 1.51 |\n| 4 | 7.0 | 38.51 | 38.95 | 42.55 | 37.98 | 86.62 | 87.08 | 87.50 | 86.20 | 2.27 |\n| 8 | 7.0 | 42.82 | 43.00 | 47.11 | 42.55 | 122.05 | 122.29 | 122.70 | 121.59 | 2.86 |\n| 16 | 7.0 | 67.74 | 67.92 | 69.05 | 65.69 | 149.92 | 150.16 | 151.03 | 149.49 | 2.28 |\n| 1 | 16.7 | 39.28 | 39.78 | 43.34 | 38.35 | 66.73 | 67.16 | 69.80 | 66.01 | 1.72 |\n| 2 | 16.7 | 43.05 | 43.42 | 47.18 | 42.43 | 120.04 | 121.12 | 123.32 | 118.14 | 2.78 |\n| 4 | 16.7 | 52.18 | 52.49 | 56.11 | 51.63 | 176.09 | 176.51 | 178.70 | 174.60 | 3.38 |\n| 8 | 16.7 | 78.55 | 78.79 | 81.66 | 78.04 | 216.19 | 216.68 | 217.63 | 214.48 | 2.75 |\n| 16 | 16.7 | 125.57 | 125.92 | 128.78 | 124.33 | 264.11 | 264.49 | 266.14 | 262.80 | 2.11 |\n\nTo achieve these same results, follow the [Quick Start Guide](#quick-start-guide) outlined above.\n\n##### Inference performance: NVIDIA T4\n\n| | | FP16 Latency (ms) Percentiles | | | | FP32 Latency (ms) Percentiles | | | | FP16/FP32 speed up |\n|-----:|---------------:|-------:|-------:|-------:|-------:|-------:|-------:|-------:|-------:|------:|\n| BS | Duration (s) | 90% | 95% | 99% | Avg | 90% | 95% | 99% | Avg | Avg |\n| 1 | 2.0 | 43.62 | 46.95 | 50.46 | 37.23 | 51.31 | 52.37 | 56.21 | 49.77 | 1.34 |\n| 2 | 2.0 | 49.09 | 50.46 | 53.11 | 40.61 | 81.85 | 82.22 | 83.94 | 80.81 | 1.99 |\n| 4 | 2.0 | 47.71 | 51.14 | 55.09 | 41.29 | 112.56 | 115.13 | 118.56 | 111.60 | 2.70 |\n| 8 | 2.0 | 51.37 | 53.11 | 55.48 | 45.94 | 198.95 | 199.48 | 200.28 | 197.22 | 4.29 |\n| 16 | 2.0 | 63.59 | 64.30 | 66.90 | 61.77 | 221.75 | 222.07 | 223.22 | 220.09 | 3.56 |\n| 1 | 7.0 | 47.49 | 48.66 | 53.36 | 40.76 | 73.63 | 74.41 | 77.65 | 72.41 | 1.78 |\n| 2 | 7.0 | 48.63 | 50.01 | 58.35 | 43.44 | 114.66 | 115.28 | 117.63 | 112.41 | 2.59 |\n| 4 | 7.0 | 52.19 | 52.85 | 54.22 | 49.94 | 200.38 | 201.29 | 202.97 | 197.21 | 3.95 |\n| 8 | 7.0 | 84.90 | 85.56 | 87.52 | 83.41 | 404.00 | 404.72 | 405.70 | 400.25 | 4.80 |\n| 16 | 7.0 | 157.12 | 157.58 | 159.19 | 155.01 | 490.93 | 492.09 | 493.44 | 486.45 | 3.14 |\n| 1 | 16.7 | 50.57 | 51.57 | 57.58 | 46.27 | 150.39 | 151.84 | 153.54 | 147.31 | 3.18 |\n| 2 | 16.7 | 63.64 | 64.55 | 66.31 | 61.98 | 256.54 | 258.16 | 262.71 | 250.34 | 4.04 |\n| 4 | 16.7 | 140.44 | 141.06 | 142.00 | 138.14 | 519.59 | 521.41 | 523.86 | 512.74 | 3.71 |\n| 8 | 16.7 | 267.03 | 268.06 | 270.01 | 263.15 | 727.33 | 728.61 | 731.36 | 722.62 | 2.75 |\n| 16 | 16.7 | 362.40 | 364.02 | 367.80 | 358.75 | 867.92 | 869.19 | 871.46 | 860.37 | 2.40 |\n\nTo achieve these same results, follow the [Quick Start Guide](#quick-start-guide) outlined above.","hasPerformance",[],{"_86":183,"_113":235,"_114":-5,"_122":34,"_121":34,"_123":234,"_218":91},{"_86":28,"_113":28,"_114":-5,"_122":34,"_121":34,"_123":234,"_218":91},[],"Latest (20.10.0)","nvidia/dle/jasper_pyt","key","nvidia/dle/resources/jasper_pyt","resourceId","jasper_pyt","shortDescription","The Jasper model is an end-to-end neural acoustic model for automatic speech recognition (ASR).","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/jasper-pyt)\n\nThis repository provides an implementation of the Jasper model in PyTorch from the paper `Jasper: An End-to-End Convolutional Neural Acoustic Model` [https://arxiv.org/pdf/1904.03288.pdf](https://arxiv.org/pdf/1904.03288.pdf).\nThe Jasper model is an end-to-end neural acoustic model for automatic speech recognition (ASR) that provides near state-of-the-art results on LibriSpeech among end-to-end ASR models without any external data. The Jasper architecture of convolutional layers was designed to facilitate fast GPU inference, by allowing whole sub-blocks to be fused into a single GPU kernel. This is important for meeting strict real-time requirements of ASR systems in deployment.\n\nThe results of the acoustic model are combined with the results of external language models to get the top-ranked word sequences\ncorresponding to a given audio segment. This post-processing step is called decoding.\n\nThis repository is a PyTorch implementation of Jasper and provides scripts to train the Jasper 10x5 model with dense residuals from scratch on the [Librispeech](http://www.openslr.org/12) dataset to achieve the greedy decoding results of the original paper.\nThe original reference code provides Jasper as part of a research toolkit in TensorFlow [openseq2seq](https://github.com/NVIDIA/OpenSeq2Seq).\nThis repository provides a simple implementation of Jasper with scripts for training and replicating the Jasper paper results.\nThis includes data preparation scripts, training and inference scripts.\nBoth training and inference scripts offer the option to use Automatic Mixed Precision (AMP) to benefit from Tensor Cores for better performance.\n\nIn addition to providing the hyperparameters for training a model checkpoint, we publish a thorough inference analysis across different NVIDIA GPU platforms, for example, DGX A100, DGX-1, DGX-2 and T4.\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 3x 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\nThe original paper takes the output of the Jasper acoustic model and shows results for 3 different decoding variations: greedy decoding, beam search with a 6-gram language model and beam search with further rescoring of the best ranked hypotheses with Transformer XL, which is a neural language model. Beam search and the rescoring with the neural language model scores are run on CPU and result in better word error rates compared to greedy decoding.\nThis repository provides instructions to reproduce greedy decoding results. To run beam search or rescoring with TransformerXL, use the following scripts from the [openseq2seq](https://github.com/NVIDIA/OpenSeq2Seq) repository:\nhttps://github.com/NVIDIA/OpenSeq2Seq/blob/master/scripts/decode.py\nhttps://github.com/NVIDIA/OpenSeq2Seq/tree/master/external_lm_rescore\n\n### Model architecture\nDetails on the model architecture can be found in the paper [Jasper: An End-to-End Convolutional Neural Acoustic Model](https://arxiv.org/pdf/1904.03288.pdf).\n\n|

|\n|:---:|:---:|\n|Figure 1: Jasper BxR model: B- number of blocks, R- number of sub-blocks | Figure 2: Jasper Dense Residual |\n\nJasper is an end-to-end neural acoustic model that is based on convolutions.\nIn the audio processing stage, each frame is transformed into mel-scale spectrogram features, which the acoustic model takes as input and outputs a probability distribution over the vocabulary for each frame.\nThe acoustic model has a modular block structure and can be parametrized accordingly:\na Jasper BxR model has B blocks, each consisting of R repeating sub-blocks.\n\nEach sub-block applies the following operations in sequence: 1D-Convolution, Batch Normalization, ReLU activation, and Dropout.\n\nEach block input is connected directly to the last subblock of all following blocks via a residual connection, which is referred to as `dense residual` in the paper.\nEvery block differs in kernel size and number of filters, which are increasing in size from the bottom to the top layers.\nIrrespective of the exact block configuration parameters B and R, every Jasper model has four additional convolutional blocks:\none immediately succeeding the input layer (Prologue) and three at the end of the B blocks (Epilogue).\n\nThe Prologue is to decimate the audio signal\nin time in order to process a shorter time sequence for efficiency. The Epilogue with dilation captures a bigger context around an audio time step, which decreases the model word error rate (WER).\nThe paper achieves best results with Jasper 10x5 with dense residual connections, which is also the focus of this repository and is in the following referred to as Jasper Large.\n\n### Default configuration\nThe following features were implemented in this model:\n\n* GPU-supported feature extraction with data augmentation options [SpecAugment](https://arxiv.org/abs/1904.08779) and [Cutout](https://arxiv.org/pdf/1708.04552.pdf)\n* offline and online [Speed Perturbation](https://www.danielpovey.com/files/2015_interspeech_augmentation.pdf)\n* data-parallel multi-GPU training and evaluation\n* AMP with dynamic loss scaling for Tensor Core training\n* FP16 inference\n\nCompetitive training results and analysis is provided for the following Jasper model configuration\n\n| **Model** | **Number of Blocks** | **Number of Subblocks** | **Max sequence length** | **Number of Parameters** |\n|--------------|----------------------|-------------------------|-------------------------|--------------------------|\n| Jasper Large | 10 | 5 | 16.7 s | 333 M |\n\n### Feature support matrix\nThe following features are supported by this model.\n\n| **Feature** | **Jasper** |\n|---------------|---------------|\n|[Apex AMP](https://nvidia.github.io/apex/amp.html) | Yes |\n|[Apex DistributedDataParallel](https://nvidia.github.io/apex/parallel.html#apex.parallel.DistributedDataParallel) | Yes |\n\n#### Features\n[Apex AMP](https://nvidia.github.io/apex/amp.html) - a tool that enables Tensor Core-accelerated training. Refer to the [Enabling mixed precision](#enabling-mixed-precision) section for more details.\n\n[Apex\nDistributedDataParallel](https://nvidia.github.io/apex/parallel.html#apex.parallel.DistributedDataParallel) -\na module wrapper that enables easy multiprocess distributed data parallel\ntraining, similar to\n[torch.nn.parallel.DistributedDataParallel](https://pytorch.org/docs/stable/nn.html#torch.nn.parallel.DistributedDataParallel).\n`DistributedDataParallel` is optimized for use with\n[NCCL](https://github.com/NVIDIA/nccl). It achieves high performance by\noverlapping communication with computation during `backward()` and bucketing\nsmaller gradient transfers to reduce the total number of transfers required.\n\n### Mixed precision training\n*Mixed 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 half-precision format, while storing minimal information in single-precision to retain 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 overall speedup on the most arithmetically 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\nThe ability to train deep learning networks with lower precision was introduced in the Pascal architecture and first supported in [CUDA 8](https://devblogs.nvidia.com/parallelforall/tag/fp16/) in the NVIDIA Deep Learning SDK.\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/sdk/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* APEX tools for mixed precision training, see the [NVIDIA Apex: Tools for Easy Mixed-Precision Training in PyTorch](https://devblogs.nvidia.com/apex-pytorch-easy-mixed-precision-training/).\n\n#### Enabling mixed precision\nFor training, mixed precision can be enabled by setting the flag: `train.py --amp`. When using bash helper scripts: `scripts/train.sh` `scripts/inference.sh`, etc., mixed precision can be enabled with env variable `AMP=true`.\n\nMixed precision is enabled in PyTorch by using the Automatic Mixed Precision\n(AMP) library from [APEX](https://github.com/NVIDIA/apex) that casts variables\nto half-precision upon retrieval, while storing variables in single-precision\nformat. Furthermore, to preserve small gradient magnitudes in backpropagation,\na [loss\nscaling](https://docs.nvidia.com/deeplearning/sdk/mixed-precision-training/index.html#lossscaling)\nstep must be included when applying gradients. In PyTorch, loss scaling can be\neasily applied by using `scale_loss()` method provided by AMP. The scaling\nvalue to be used can be\n[dynamic](https://nvidia.github.io/apex/amp.html#apex.amp.initialize) or fixed.\n\nFor an in-depth walk through on AMP, check out sample usage\n[here](https://nvidia.github.io/apex/amp.html#). [APEX](https://github.com/NVIDIA/apex) is a PyTorch extension that contains\nutility libraries, such as AMP, which require minimal network code changes to\nleverage Tensor Cores performance.\n\nThe following steps were needed to enable mixed precision training in Jasper:\n\n* Import AMP from APEX (file: `train.py`):\n```bash\nfrom apex import amp\n```\n\n* Initialize AMP and wrap the model and the optimizer\n```bash\n model, optimizer = amp.initialize(\n min_loss_scale=1.0,\n models=model,\n optimizers=optimizer,\n opt_level='O1')\n\n```\n\n* Apply `scale_loss` context manager\n```bash\nwith amp.scale_loss(loss, optimizer) as scaled_loss:\n scaled_loss.backward()\n```\n\n#### Enabling TF32\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 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 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**Acoustic model**\nAssigns a probability distribution over a vocabulary of characters given an audio frame.\n\n**Language Model**\nAssigns a probability distribution over a sequence of words. Given a sequence of words, it assigns a probability to the whole sequence.\n\n**Pre-training**\nTraining a model on vast amounts of data on the same (or different) task to build general understandings.\n\n**Automatic Speech Recognition (ASR)**\nUses both acoustic model and language model to output the transcript of an input audio signal.","displayName","Jasper for PyTorch",["D",1680632392117],"publisher","NVIDIA Deep Learning Examples","latestVersion","collectionCounts","imageUrl","/api/image/nvidia/dle/resources/jasper_pyt","rawImageUrl","https://assets.nvidiagrid.net/ngc/logos/Deep-Learning-Examples.png","canGuestDownload","latestVersionSizeInBytes","licenseTerms","isMultinodeEnabled","isMultiArchitecture","bias","explainability","privacy","safetyAndSecurity","featureLabels",[],"displayLabels",[275],"policyLabels",[],"systemLabels",[],"products",[],"hasNimMetadata",{"_244":276,"_277":278,"_279":280,"_281":88},"Natural Language Understanding","value","uscs_natural_language_understanding","filter",{"_282":283,"_277":278},"category","field","application",{"_152":1261,"_154":1262},"ngcApps",[1214],"csaApiBaseUrl","https://www.nvidia.com/ai-agents/csa-agent/dcsm","datadogRum",{"_1195":1196,"_1197":1198,"_1199":1200,"_1201":1202,"_1203":295,"_110":321,"_1204":1205,"_1206":59,"_1207":91,"_1208":91,"_1209":1210,"_1211":91,"_1212":1213},"amplitude",{"_1192":1193,"_833":-7,"_1194":34},"disableJs","environment","prod","featureFlagsBlob","JTdCJTIydGVzdEZsYWclMjIlM0FmYWxzZSUyQyUyMnNob3dTZWFyY2hMYWJlbFNldEZpbHRlcnMlMjIlM0FmYWxzZSUyQyUyMmVuYWJsZUNzYVdpZGdldCUyMiUzQWZhbHNlJTJDJTIyZW5hYmxlUHJvZHVjdEdyb3VwaW5nQnlTdWJzY3JpcHRpb24lMjIlM0FmYWxzZSUyQyUyMmNvbnRhaW5lclNjYW5uaW5nUmVkZXNpZ25NMiUyMiUzQWZhbHNlJTdE","featureFlagsGuarded","isDevLikeEnv","isPlaywright","isMobileDevice","categorizedOrgs",{"_1186":1187,"_1188":1189,"_1190":1191},"currentOrgContext","currentContextRoles",[],"enterpriseOrgs",[],"manageableExternalUserGroups",[],"ngcApiOrigin","ngcHelmOrigin","origin","https://catalog.ngc.nvidia.com","usesMocks","theme","year",2026,"now",1787560659272,"1.12.0","defaultSearchSuggestions",{"_1073":401,"_1074":1075,"_1076":1077,"_1078":1079},"announcements",[],"consentState",{"_358":1063,"_1064":91,"_1065":34,"_1066":34,"_1067":1068},"consentConfig",{"_340":341,"_342":343,"_344":345,"_346":347,"_348":349,"_350":351,"_352":353,"_354":355,"_356":357,"_358":359},"showContactAdmin",1787560659273,[],[],[335,336,337,338,339],{"_48":63,"_50":74,"_52":147,"_54":-5,"_56":-5,"_58":59},{"_48":63,"_50":70,"_52":147,"_54":-5,"_56":-5,"_58":59},{"_48":63,"_50":64,"_52":147,"_54":-5,"_56":-5,"_58":59},{"_48":60,"_50":61,"_52":53,"_54":55,"_56":57,"_58":59},{"_48":49,"_50":51,"_52":147,"_54":-5,"_56":-5,"_58":59},"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",{"_1055":1056,"_1057":1058,"_1059":1060,"_1061":1062},"bannerAcceptLabel","Accept All","bannerRejectLabel","Reject Optional","bannerSettingsLabel","Manage Settings","preferenceSaveLabel","Save Choices","preferenceRejectLabel","Decline All","alwaysActiveText","Always Active","groups",[360,361,362,363,364],{"_365":745,"_113":746,"_212":747,"_369":91,"_370":16,"_372":748,"_374":749},{"_365":648,"_113":649,"_212":650,"_369":34,"_370":57,"_372":651,"_374":652},{"_365":556,"_113":557,"_212":558,"_369":34,"_370":62,"_372":559,"_374":560},{"_365":376,"_113":377,"_212":378,"_369":34,"_370":379,"_372":380,"_374":381},{"_365":366,"_113":367,"_212":368,"_369":34,"_370":371,"_372":373,"_374":375},"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,[498,499,500,501,502,503,504,505,506,507,508,509,510,511,512,513,514,515,516,517,518,519],[382,383,384,385,386,387,388,389,390,391,392,393,394,395,396,397,398],{"_399":493,"_244":493,"_212":401,"_372":494},{"_399":489,"_244":489,"_212":401,"_372":490},{"_399":484,"_244":484,"_212":401,"_372":485},{"_399":480,"_244":480,"_212":401,"_372":481},{"_399":475,"_244":475,"_212":401,"_372":476},{"_399":470,"_244":470,"_212":401,"_372":471},{"_399":462,"_244":462,"_212":401,"_372":463},{"_399":454,"_244":454,"_212":401,"_372":455},{"_399":449,"_244":449,"_212":401,"_372":450},{"_399":444,"_244":444,"_212":401,"_372":445},{"_399":439,"_244":439,"_212":401,"_372":440},{"_399":434,"_244":434,"_212":401,"_372":435},{"_399":429,"_244":429,"_212":401,"_372":430},{"_399":422,"_244":422,"_212":401,"_372":423},{"_399":417,"_244":417,"_212":401,"_372":418},{"_399":410,"_244":410,"_212":401,"_372":411},{"_399":400,"_244":400,"_212":401,"_372":402},"hostName","amazon-adsystem.com","",[403,404],{"_113":409,"_406":400,"_407":34,"_212":408},{"_113":405,"_406":400,"_407":34,"_212":408},"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",[412,413],{"_113":416,"_406":410,"_407":34,"_212":415},{"_113":414,"_406":410,"_407":34,"_212":415},"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",[419],{"_113":420,"_406":417,"_407":34,"_212":421},"MUID","This domain is owned by Mircosoft - it is the site for the search engine Bing.","assets.adobedtm.com",[424,425],{"_113":428,"_406":422,"_407":34,"_212":427},{"_113":426,"_406":422,"_407":34,"_212":427},"kndctr_F207D74D549850760A4C98C6_AdobeOrg_*","If you know anything about this host and its uses please add your contributions here. ","kndctr_F207D74D549850760A4C98C6_AdobeOrg_consent","tiktok.com",[431],{"_113":432,"_406":429,"_407":34,"_212":433},"_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",[436],{"_113":437,"_406":434,"_407":34,"_212":438},"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",[441],{"_113":442,"_406":439,"_407":34,"_212":443},"sp","Testing","google.com",[446],{"_113":447,"_406":444,"_407":34,"_212":448},"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",[451],{"_113":452,"_406":449,"_407":34,"_212":453},"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",[456,457],{"_113":460,"_406":454,"_407":34,"_212":461},{"_113":458,"_406":454,"_407":34,"_212":459},"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",[464,465],{"_113":468,"_406":462,"_407":34,"_212":469},{"_113":466,"_406":462,"_407":34,"_212":467},"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",[472],{"_113":473,"_406":470,"_407":34,"_212":474},"demdex","This cookie helps Adobe Audience Manger perform basic functions such as visitor identification, ID synchronization, segmentation, modeling, reporting, etc.","im-apps.net",[477],{"_113":478,"_406":475,"_407":34,"_212":479},"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",[482],{"_113":483,"_406":480,"_407":34,"_212":443},"R","6sc.co",[486],{"_113":487,"_406":484,"_407":34,"_212":488},"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",[491],{"_113":492,"_406":489,"_407":34,"_212":443},"_hly_gvid","t.contentsquare.net",[495],{"_113":496,"_406":493,"_407":91,"_212":497},"_cs_*","This is a session cookie. This is a pattern type cookie where a random 13-digit number is appended to the prefix _cs_.",{"_113":554,"_406":521,"_407":34,"_212":555},{"_113":552,"_406":521,"_407":34,"_212":553},{"_113":550,"_406":521,"_407":34,"_212":551},{"_113":432,"_406":521,"_407":34,"_212":549},{"_113":526,"_406":521,"_407":91,"_212":528},{"_113":548,"_406":521,"_407":34,"_212":401},{"_113":547,"_406":524,"_407":34,"_212":401},{"_113":545,"_406":521,"_407":34,"_212":546},{"_113":543,"_406":521,"_407":34,"_212":544},{"_113":542,"_406":521,"_407":34,"_212":401},{"_113":540,"_406":521,"_407":34,"_212":541},{"_113":526,"_406":539,"_407":91,"_212":528},{"_113":537,"_406":524,"_407":34,"_212":538},{"_113":537,"_406":527,"_407":34,"_212":538},{"_113":496,"_406":521,"_407":34,"_212":497},{"_113":535,"_406":521,"_407":34,"_212":536},{"_113":442,"_406":533,"_407":34,"_212":534},{"_113":531,"_406":521,"_407":34,"_212":532},{"_113":529,"_406":521,"_407":34,"_212":530},{"_113":526,"_406":527,"_407":91,"_212":528},{"_113":523,"_406":524,"_407":34,"_212":525},{"_113":520,"_406":521,"_407":91,"_212":522},"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.",[590,591,592,593,594,595,596,597,598,599,600,601,602,603,604,605,606,607,608,609,610,611,612,613,614],[561,562,563,564,565],{"_399":475,"_244":475,"_212":401,"_372":585},{"_399":449,"_244":449,"_212":401,"_372":580},{"_399":575,"_244":575,"_212":401,"_372":576},{"_399":570,"_244":570,"_212":401,"_372":571},{"_399":566,"_244":566,"_212":401,"_372":567},"nvidia.partners",[568],{"_113":569,"_406":566,"_407":34,"_212":443},"edge","app.hushly.com",[572],{"_113":573,"_406":570,"_407":34,"_212":574},"*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",[577],{"_113":578,"_406":575,"_407":34,"_212":579},"_octo","This domain is owned by Github. The main business activity is: Source code repository and collaboration platform",[581,582],{"_113":584,"_406":449,"_407":34,"_212":453},{"_113":583,"_406":449,"_407":34,"_212":453},"lidc","li_gc",[586,587],{"_113":589,"_406":475,"_407":34,"_212":479},{"_113":588,"_406":475,"_407":34,"_212":479},"imid_*","imid",{"_113":615,"_406":524,"_407":34,"_212":617},{"_113":646,"_406":521,"_407":34,"_212":647},{"_113":615,"_406":645,"_407":34,"_212":617},{"_113":643,"_406":521,"_407":91,"_212":644},{"_113":641,"_406":521,"_407":34,"_212":642},{"_113":640,"_406":623,"_407":91,"_212":401},{"_113":615,"_406":619,"_407":34,"_212":617},{"_113":638,"_406":521,"_407":34,"_212":639},{"_113":615,"_406":527,"_407":34,"_212":617},{"_113":626,"_406":524,"_407":34,"_212":401},{"_113":615,"_406":637,"_407":34,"_212":617},{"_113":635,"_406":521,"_407":34,"_212":636},{"_113":632,"_406":633,"_407":34,"_212":634},{"_113":630,"_406":631,"_407":34,"_212":401},{"_113":615,"_406":627,"_407":34,"_212":617},{"_113":629,"_406":521,"_407":34,"_212":401},{"_113":615,"_406":628,"_407":34,"_212":617},{"_113":626,"_406":627,"_407":34,"_212":401},{"_113":625,"_406":623,"_407":91,"_212":401},{"_113":615,"_406":624,"_407":34,"_212":617},{"_113":615,"_406":623,"_407":34,"_212":617},{"_113":615,"_406":622,"_407":34,"_212":617},{"_113":621,"_406":521,"_407":91,"_212":617},{"_113":618,"_406":619,"_407":91,"_212":620},{"_113":615,"_406":616,"_407":34,"_212":617},"_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.",[673,674,675,676,677,678,679,680,681,682,683,684,685,686,687,688,689,690,691,692,693,694,695,696,697,698,699,700,701],[653,654,655,656],{"_399":668,"_244":668,"_212":401,"_372":669},{"_399":663,"_244":663,"_212":401,"_372":664},{"_399":462,"_244":462,"_212":401,"_372":660},{"_399":575,"_244":575,"_212":401,"_372":657},[658],{"_113":659,"_406":575,"_407":91,"_212":579},"_gh_sess",[661],{"_113":662,"_406":462,"_407":91,"_212":469},"YSC","bizible.com",[665],{"_113":666,"_406":663,"_407":34,"_212":667},"_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",[670],{"_113":671,"_406":668,"_407":34,"_212":672},"tuid","This domain is owned by Ookla which provides internet speed testing services through its Speedtest application.",{"_113":743,"_406":521,"_407":34,"_212":744},{"_113":728,"_406":742,"_407":91,"_212":729},{"_113":711,"_406":524,"_407":34,"_212":636},{"_113":711,"_406":624,"_407":34,"_212":636},{"_113":740,"_406":521,"_407":34,"_212":741},{"_113":728,"_406":527,"_407":91,"_212":729},{"_113":738,"_406":521,"_407":34,"_212":739},{"_113":728,"_406":524,"_407":91,"_212":729},{"_113":736,"_406":623,"_407":34,"_212":737},{"_113":728,"_406":645,"_407":91,"_212":729},{"_113":708,"_406":524,"_407":34,"_212":710},{"_113":734,"_406":521,"_407":91,"_212":735},{"_113":732,"_406":521,"_407":34,"_212":733},{"_113":730,"_406":521,"_407":91,"_212":731},{"_113":728,"_406":619,"_407":91,"_212":729},{"_113":726,"_406":521,"_407":34,"_212":727},{"_113":724,"_406":521,"_407":34,"_212":725},{"_113":722,"_406":521,"_407":91,"_212":723},{"_113":720,"_406":628,"_407":34,"_212":721},{"_113":708,"_406":619,"_407":34,"_212":710},{"_113":718,"_406":521,"_407":34,"_212":719},{"_113":716,"_406":521,"_407":91,"_212":717},{"_113":714,"_406":521,"_407":34,"_212":715},{"_113":712,"_406":521,"_407":34,"_212":713},{"_113":711,"_406":521,"_407":34,"_212":636},{"_113":708,"_406":709,"_407":34,"_212":710},{"_113":706,"_406":521,"_407":34,"_212":707},{"_113":704,"_406":521,"_407":91,"_212":705},{"_113":702,"_406":521,"_407":34,"_212":703},"_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.",[890,891,892,893,894,895,896,897,898,899,900,901,902,903,904,905,906,907,908,909,910,911,912,913,914,915,916,917,918,919,920,921,922,923,924,925,926,927,928,929,930,931,932,933,934,935,936,937,938,939,940,941,942,943,944,945,946,947,948,949,950,951,952,953,954,955,956,957,958,959,960,961,962],[750,751,752,753,754,755,756,757,758,759,760,761,762,763,764,765,766,767,768,769,770,771,772,773,774],{"_399":887,"_244":887,"_212":401,"_372":888},{"_399":881,"_244":881,"_212":401,"_372":882},{"_399":878,"_244":878,"_212":401,"_372":879},{"_399":866,"_244":866,"_212":401,"_372":867},{"_399":863,"_244":863,"_212":401,"_372":864},{"_399":860,"_244":860,"_212":401,"_372":861},{"_399":855,"_244":855,"_212":401,"_372":856},{"_399":462,"_244":462,"_212":401,"_372":852},{"_399":849,"_244":849,"_212":401,"_372":850},{"_399":454,"_244":454,"_212":401,"_372":847},{"_399":449,"_244":449,"_212":401,"_372":845},{"_399":840,"_244":840,"_212":401,"_372":841},{"_399":835,"_244":835,"_212":401,"_372":836},{"_399":830,"_244":830,"_212":401,"_372":831},{"_399":827,"_244":827,"_212":401,"_372":828},{"_399":575,"_244":575,"_212":401,"_372":824},{"_399":434,"_244":434,"_212":401,"_372":821},{"_399":570,"_244":570,"_212":401,"_372":814},{"_399":566,"_244":566,"_212":401,"_372":808},{"_399":802,"_244":802,"_212":401,"_372":803},{"_399":797,"_244":797,"_212":401,"_372":798},{"_399":792,"_244":792,"_212":401,"_372":793},{"_399":788,"_244":788,"_212":401,"_372":789},{"_399":783,"_244":783,"_212":401,"_372":784},{"_399":775,"_244":775,"_212":401,"_372":776},"polyfill.io",[777,778],{"_113":781,"_406":775,"_407":34,"_212":782},{"_113":779,"_406":775,"_407":34,"_212":780},"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",[785],{"_113":786,"_406":783,"_407":34,"_212":787},"_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",[790],{"_113":791,"_406":788,"_407":34,"_212":443},"__cflb","instagram.com",[794],{"_113":795,"_406":792,"_407":34,"_212":796},"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",[799],{"_113":800,"_406":797,"_407":34,"_212":801},"__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",[804,805],{"_113":807,"_406":802,"_407":34,"_212":443},{"_113":806,"_406":802,"_407":34,"_212":443},"AWSALBTG","AWSALBTGCORS",[809,810],{"_113":812,"_406":566,"_407":34,"_212":813},{"_113":811,"_406":566,"_407":34,"_212":443},"ak_bmsc","bm_mi","This domain is likely associated with NVIDIA's partner program, which could involve marketing, sales, and support services for partners.",[815,816,817],{"_113":820,"_406":570,"_407":34,"_212":443},{"_113":819,"_406":570,"_407":91,"_212":443},{"_113":818,"_406":570,"_407":34,"_212":443},"AWSALBCORS","JSESSIONID","AWSALB",[822],{"_113":800,"_406":434,"_407":34,"_212":823},"This is a CloudFoundry cookie",[825],{"_113":826,"_406":575,"_407":34,"_212":579},"logged_in","hcaptcha.com",[829],{"_113":800,"_406":827,"_407":34,"_212":823},"tracking.intentsify.io",[832],{"_113":833,"_406":830,"_407":34,"_212":834},"userId","Used to track usage of the embedded video player.","static.reo.dev",[837],{"_113":838,"_406":835,"_407":34,"_212":839},"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",[842],{"_113":843,"_406":840,"_407":34,"_212":844},"dd_cookie_test_","dd_cookie_test",[846],{"_113":800,"_406":449,"_407":34,"_212":823},[848],{"_113":800,"_406":454,"_407":34,"_212":823},"supabase.co",[851],{"_113":800,"_406":849,"_407":34,"_212":823},[853],{"_113":854,"_406":462,"_407":34,"_212":469},"VISITOR_PRIVACY_METADATA","dosa-prd.azurewebsites.net",[857],{"_113":858,"_406":855,"_407":34,"_212":859},"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",[862],{"_113":800,"_406":860,"_407":34,"_212":823},"cdn.readme.io",[865],{"_113":800,"_406":863,"_407":34,"_212":823},"queue-it.net",[868,869,870,871,872],{"_113":877,"_406":866,"_407":34,"_212":401},{"_113":876,"_406":866,"_407":34,"_212":401},{"_113":875,"_406":866,"_407":34,"_212":401},{"_113":874,"_406":866,"_407":34,"_212":401},{"_113":873,"_406":866,"_407":34,"_212":401},"Queue-it-{queueId}","Queue-it-{dmonvidia }________{eventId}","Queue-it-token","queueitsoftblock_{queueId}","Queue-it-{nvidia}________{eventId}","x.com",[880],{"_113":800,"_406":878,"_407":34,"_212":801},"api.hcaptcha.com",[883,884],{"_113":886,"_406":881,"_407":34,"_212":443},{"_113":791,"_406":881,"_407":34,"_212":885},"This domain is associated with hCaptcha, a service that provides CAPTCHA and reCAPTCHA alternatives for user verification and bot protection.","hmt_id","vimeo.com",[889],{"_113":800,"_406":887,"_407":34,"_212":823},{"_113":1023,"_406":521,"_407":34,"_212":1054},{"_113":1052,"_406":616,"_407":91,"_212":1053},{"_113":1045,"_406":1014,"_407":91,"_212":1046},{"_113":1051,"_406":709,"_407":91,"_212":401},{"_113":1049,"_406":521,"_407":34,"_212":1050},{"_113":807,"_406":1035,"_407":34,"_212":974},{"_113":982,"_406":1044,"_407":91,"_212":984},{"_113":806,"_406":973,"_407":34,"_212":974},{"_113":985,"_406":1044,"_407":34,"_212":986},{"_113":1047,"_406":521,"_407":91,"_212":1048},{"_113":1045,"_406":1021,"_407":91,"_212":1046},{"_113":979,"_406":1044,"_407":34,"_212":981},{"_113":1042,"_406":521,"_407":34,"_212":1043},{"_113":806,"_406":965,"_407":34,"_212":974},{"_113":970,"_406":1041,"_407":34,"_212":972},{"_113":818,"_406":971,"_407":34,"_212":987},{"_113":820,"_406":971,"_407":34,"_212":1038},{"_113":985,"_406":980,"_407":34,"_212":986},{"_113":1039,"_406":619,"_407":34,"_212":1040},{"_113":820,"_406":645,"_407":34,"_212":1038},{"_113":1036,"_406":521,"_407":34,"_212":1037},{"_113":812,"_406":645,"_407":34,"_212":1031},{"_113":1008,"_406":624,"_407":91,"_212":401},{"_113":806,"_406":1035,"_407":34,"_212":974},{"_113":1033,"_406":973,"_407":91,"_212":1034},{"_113":997,"_406":521,"_407":34,"_212":988},{"_113":970,"_406":521,"_407":34,"_212":972},{"_113":1032,"_406":533,"_407":34,"_212":401},{"_113":812,"_406":521,"_407":34,"_212":1031},{"_113":820,"_406":965,"_407":34,"_212":1030},{"_113":1028,"_406":628,"_407":91,"_212":1029},{"_113":1026,"_406":521,"_407":91,"_212":1027},{"_113":1025,"_406":521,"_407":34,"_212":401},{"_113":1023,"_406":971,"_407":34,"_212":1024},{"_113":1022,"_406":533,"_407":34,"_212":401},{"_113":1013,"_406":1021,"_407":91,"_212":1015},{"_113":1019,"_406":624,"_407":34,"_212":1020},{"_113":1017,"_406":709,"_407":34,"_212":1018},{"_113":1016,"_406":527,"_407":34,"_212":401},{"_113":1013,"_406":1014,"_407":91,"_212":1015},{"_113":1011,"_406":980,"_407":91,"_212":1012},{"_113":1009,"_406":521,"_407":34,"_212":1010},{"_113":1008,"_406":524,"_407":91,"_212":401},{"_113":979,"_406":983,"_407":34,"_212":981},{"_113":811,"_406":998,"_407":34,"_212":988},{"_113":1006,"_406":524,"_407":34,"_212":1007},{"_113":1004,"_406":521,"_407":34,"_212":1005},{"_113":1002,"_406":521,"_407":34,"_212":1003},{"_113":1001,"_406":521,"_407":34,"_212":401},{"_113":843,"_406":524,"_407":34,"_212":1000},{"_113":997,"_406":645,"_407":34,"_212":988},{"_113":800,"_406":616,"_407":34,"_212":999},{"_113":997,"_406":998,"_407":34,"_212":988},{"_113":996,"_406":533,"_407":34,"_212":401},{"_113":995,"_406":521,"_407":34,"_212":968},{"_113":993,"_406":980,"_407":34,"_212":994},{"_113":806,"_406":991,"_407":34,"_212":992},{"_113":990,"_406":533,"_407":34,"_212":401},{"_113":989,"_406":521,"_407":34,"_212":401},{"_113":982,"_406":980,"_407":91,"_212":984},{"_113":811,"_406":645,"_407":34,"_212":988},{"_113":818,"_406":645,"_407":34,"_212":987},{"_113":985,"_406":983,"_407":34,"_212":986},{"_113":982,"_406":983,"_407":91,"_212":984},{"_113":979,"_406":980,"_407":34,"_212":981},{"_113":977,"_406":973,"_407":91,"_212":978},{"_113":975,"_406":631,"_407":34,"_212":976},{"_113":807,"_406":973,"_407":34,"_212":974},{"_113":970,"_406":971,"_407":34,"_212":972},{"_113":800,"_406":623,"_407":34,"_212":969},{"_113":967,"_406":521,"_407":34,"_212":968},{"_113":818,"_406":965,"_407":34,"_212":966},{"_113":963,"_406":521,"_407":34,"_212":964},"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.

",{"_745":16,"_648":59,"_556":59,"_376":59,"_366":59},"hasConsented","gpcApplied","showBanner","geolocation",{"_1069":1070,"_1071":1072},"country","US","state","OR","query","categories",[1092,1093,1094,1095,1096],"optionUrls",[1080,1081,1082,1083,1084,1085,1086,1087,1088,1089,1090,1091],"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/-",{"_20":1170,"_1098":1171,"_1100":59},{"_20":1135,"_1098":1136,"_1100":57},{"_20":1118,"_1098":1119,"_1100":1120},{"_20":21,"_1098":1111,"_1100":1112},{"_20":1097,"_1098":1099,"_1100":19},"helm-charts","items",[1101,1102],"startIndex",{"_20":1097,"_244":1108,"_206":207,"_208":-7,"_239":1109,"_247":1105,"_272":1110,"_1107":1090},{"_20":1097,"_244":1103,"_206":207,"_208":-7,"_239":1104,"_247":1105,"_272":1106,"_1107":1091},"NVIDIA NIM Operator","k8s-nim-operator","NVIDIA",[],"url","GPU Operator","gpu-operator",[],[1113],9,{"_20":21,"_244":1114,"_206":207,"_208":1115,"_239":1116,"_247":1105,"_272":1117,"_1107":1089},"Riva Skills Quick Start","riva","riva_quickstart",[],"models",[1121,1122,1123],6,{"_20":1118,"_244":1131,"_206":207,"_208":1132,"_239":1133,"_247":1105,"_272":1134,"_1107":1086},{"_20":1118,"_244":1128,"_206":207,"_208":1125,"_239":1129,"_247":1105,"_272":1130,"_1107":1087},{"_20":1118,"_244":1124,"_206":207,"_208":1125,"_239":1126,"_247":1105,"_272":1127,"_1107":1088},"TrafficCamNet","tao","trafficcamnet",[],"PeopleNet","peoplenet",[],"StyleGAN3 pretrained models","research","stylegan3",[],"collections",[1137,1138,1139],{"_20":1135,"_244":1167,"_206":207,"_208":1141,"_239":1168,"_247":1105,"_272":1169,"_1107":1083},{"_20":1135,"_244":1164,"_206":207,"_208":-7,"_239":1165,"_247":1105,"_272":1166,"_1107":1084},{"_20":1135,"_244":1140,"_206":207,"_208":1141,"_239":1142,"_247":1105,"_272":1143,"_1107":1085},"Omniverse Kit App Streaming","omniverse","kit-appstreaming-collection",[1144,1145,1146],{"_113":1158,"_244":1159,"_241":1160,"_251":1161,"_1151":1162},{"_113":1141,"_244":1154,"_241":1155,"_251":1150,"_1151":1156},{"_113":1147,"_244":1148,"_241":1149,"_251":1150,"_1151":1152},"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",[1153],"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.",[1153,1157],"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",[1153,1157,1163],"NGC_CONSOLE","DeepStream SDK","deepstream_sdk",[],"Omniverse Kit (FB)","kit",[1144,1145,1146],"containers",[1172,1173,1174],{"_20":1170,"_244":1182,"_206":1179,"_208":1183,"_239":1184,"_247":1105,"_272":1185,"_1107":1080},{"_20":1170,"_244":1178,"_206":1179,"_208":207,"_239":1180,"_247":1105,"_272":1181,"_1107":1081},{"_20":1170,"_244":1175,"_206":207,"_208":-7,"_239":1176,"_247":1105,"_272":1177,"_1107":1082},"PyTorch","pytorch",[],"Llama-3.1-Nemotron-70B-Instruct","nim","llama-3.1-nemotron-70b-instruct",[],"DeepSeek-R1","deepseek-ai","deepseek-r1",[],"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",{"_1215":1216,"_1217":314,"_1218":34,"_1219":1220,"_1098":1221,"_1222":1223,"_1224":1216,"_1225":34,"_1226":34,"_1227":34,"_1228":1229},"appName","ngc-catalog","appOrigin","hideOnPrivateOrg","icon","ObjectsModules2",[1230,1231,1232,1233,1234,1235,1236],"menuIcon","editor-swatches","productName","requiresAllScopes","requiresLogin","requiresOrgLevelScopesOnly","title","Catalog",{"_1228":1258,"_50":1259,"_1222":1260},{"_1251":1252,"_1228":1253,"_50":1254,"_1224":1158},{"_1228":1249,"_50":1250,"_1222":1223},{"_1228":1246,"_50":1247,"_1222":1248},{"_1228":1243,"_50":1244,"_1222":1245},{"_1228":1240,"_50":1241,"_1222":1242},{"_1228":1237,"_50":1238,"_1222":1239},"Resources","/resources","misc-library","Models","/models","misc-neural-network","Helm Charts","/helm-charts","misc-ship-wheel","Containers","/containers","3d-cube","Collections","/collections","access",[1255],"Enterprise Catalog","/enterprise",[1256,1257],"NVAIE","READ","Explore Catalog","/","common-home",{"_157":158,"_159":160,"_161":162,"_163":164,"_165":166,"_167":168},{"_156":34}]