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 Transformer-XL model in\nTensorFlow from the paper Transformer-XL: Attentive\nLanguage Models Beyond a Fixed-Length\nContext. Transformer-XL is a\ntransformer-based language model with a segment-level recurrence and a novel\nrelative positional encoding. Enhancements introduced in Transformer-XL help\ncapture better long-term dependencies by attending to tokens from multiple\nprevious segments.

Our implementation is based on the\ncodebase published by the\nauthors of the Transformer-XL paper.\nOur implementation uses a modified model architecture. Our\nmodifications were made to achieve better hardware utilization and to take\nadvantage of Tensor Cores. Similar modifications were also proposed in an\nimplementation available from\ngithub.com/cybertronai/transformer-xl.\nRefer to the Model architecture section for more\ndetails.

This model is trained with mixed precision using Tensor Cores on Volta, Turing,\nand the NVIDIA Ampere GPU architectures. Therefore, researchers can get\nresults up to 1.5x faster than training without Tensor Cores, while\nexperiencing the benefits of mixed precision training. This model is tested\nagainst each NGC monthly container release to ensure consistent accuracy and\nperformance over time.

Model architecture

The Transformer-XL \"base\" model for WikiText-103 dataset available in this\nrepository was modified to use the following hyperparameter values:

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

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

Hyperparameter	Description	Original setting for the base model	Our modification to the base model
`d_model`	hidden size	410	512
`n_head`	number of attention heads	10	8
`d_head`	size of each attention head	41	64
`d_inner`	hidden size in fully-connected layers	2100	2048
`tgt_len`	number of tokens to predict during training	150	192
`mem_len`	number of tokens cached from previous iterations during training	150	192

Changes described above were made to align certain hyperparameters with powers\nof two, with this modification, the model is able to achieve better hardware\nutilization, and therefore higher training throughput.

The following table lists the hyperparameters for the base\nTransformer-XL model for WikiText-103 dataset available in this repository.

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

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

Hyperparameter	Description	Base model
`n_layer`	number of layers	16
`d_model`	hidden size	512
`n_head`	number of attention heads	8
`d_head`	size of each attention head	64
`d_inner`	inner hidden size in fully-connected layers	2048
`dropout`	dropout	0.1
`dropatt`	dropout after softmax in the attention	0.0
`lr`	base learning rate	0.01
`min_lr_ratio`	minimum ratio learning rate (for cosine decay)	0.1
`max_step`	number of training steps	40,000
`warmup_step`	number of learning rate warmup steps	1,000
`batch_size`	training batch size	256
`tgt_len`	number of tokens to predict during training	192
`mem_len`	number of tokens cached from previous iterations during training	192

The Transformer-XL model addresses the limitations of vanilla transformer-based\nlanguage models, which are only able to use relatively short context, bounded\nby the segment length. The Transformer-XL introduces a recurrence mechanism,\nwhich is able to use a cached hidden state from previous segments. During\ntraining, the context consists of a concatenation of the current segment's hidden\nstate and cached states from previous iterations. Gradients are backpropagated\nonly through the current segment, although the model is able to take advantage\nof the extra information stored in the cache and therefore is able to model\nlong-term dependencies.

An illustration of the recurrence mechanism taken from the Transformer-XL\npaper is shown below.\n $\"model\"$

Default configuration

The following features were implemented in this model:

general\n
- single-node, Horovod multi-GPU training
- training and inference with mixed precision using Tensor Cores
- automatic mixed precision training (AMP)
\n
model\n
- 16-layer base Transformer-XL model with hidden size 512, 8 attention heads,\neach head with hidden size 64
- the model trained on\nWikiText-103\ndataset, using word-level vocabulary and\nadaptive softmax
- embedding weights are tied with weights in the classifier
\n
training\n
- training with LAMB optimizer, the\nimplementation of the optimizer uses XLA, which enables\nthe fusion of elementwise operations and accelerates the training
- support for training with a gradient accumulation
- base model:\n
  - linear learning rate warmup for 1,000 iterations, followed by the cosine\nlearning rate schedule, the initial learning rate is set to 0.0, and the final\nlearning rate is set to 0.001 (min_lr_ratio * base_lr)
  - training for 40,000 steps, using a batch size of 256
  \n
\n
inference\n
- support for single-GPU inference
- each token is using the same size of the context from previous time steps.
- base model:\n
  - target length is set to 64, length of memory is set to 640
  - positional embeddings are clamped after 400 time steps
  \n
\n

Feature support matrix

The following features are supported by this model:

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

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

Feature	Transformer-XL
Automatic mixed precision (AMP)	Yes
Horovod Multi-GPU (NCCL)	Yes
LAMB	Yes

Features

TF-AMP - a\ntool that enables Tensor Core-accelerated training. Refer to the Enabling\nmixed precision section for more details.

Horovod - Horovod\nis a distributed training framework for TensorFlow, Keras, PyTorch, and MXNet.\nThe goal of Horovod is to make distributed deep learning fast and easy to use.\nFor more information about how to get started with Horovod, see the Horovod:\nOfficial repository.

Multi-GPU training with Horovod - our model\nuses Horovod to implement efficient multi-GPU training with NCCL. For details,\nsee example sources in this repository or see the TensorFlow\ntutorial.

LAMB - stands\nfor Layerwise Adaptive Moments Based optimizer, is a large batch optimization\ntechnique that helps accelerate training of deep neural networks using large\nminibatches.

Mixed precision training

Mixed precision is the combined use of different numerical precisions in a\ncomputational method.\nMixed precision training offers significant\ncomputational speedup by performing operations in half-precision format while\nstoring minimal information in single-precision to retain as much information\nas possible in critical parts of the network. Since the introduction of Tensor\nCores in Volta, and following with both the\nTuring and Ampere architectures, significant training speedups are experienced by switching to\nmixed precision -- up to 3x overall speedup on the most arithmetically intense\nmodel architectures. Using mixed precision training previously required two\nsteps:

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

This can now be achieved using Automatic Mixed Precision (AMP) for TensorFlow to enable the full\nmixed precision methodology\nin your existing TensorFlow model code. AMP enables mixed precision training on Volta and Turing GPUs automatically.\nThe TensorFlow framework code makes all necessary model changes internally.

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

For information about:

How to train using mixed precision, see the Mixed Precision\nTraining paper and Training With Mixed\nPrecision\ndocumentation.
Techniques used for mixed precision training, see the Mixed-Precision\nTraining of Deep Neural\nNetworks\nblog.
How to access and enable AMP for TensorFlow, see Using\nTF-AMP\nfrom the TensorFlow User Guide.

Enabling mixed precision

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

To enable mixed precision, you can simply add the values to the environmental variables inside your training script:

Enable TF-AMP graph rewrite:

os.environ[\"TF_ENABLE_AUTO_MIXED_PRECISION_GRAPH_REWRITE\"] = \"1\"\n

Enable Automated Mixed Precision:

os.environ['TF_ENABLE_AUTO_MIXED_PRECISION'] = '1'\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.

","resourceVersions",[230,231],"resourceVersionsTotal","hasMoreVersions","versionMetadata",{"_209":28,"_210":-5,"_52":115,"_112":-5,"_116":211,"_118":-5,"_212":114,"_123":16,"_120":34,"_213":-5,"_119":34,"_214":215,"_216":89,"_217":218,"_219":89,"_220":221,"_222":89,"_223":224,"_225":89,"_226":227,"_228":89,"_121":229},"versionNotFound","isLoggedIn","latestVersionPath","/orgs/nvidia/dle/resources/transformerxl_tf1/-","-","imageReference","cliDownloadCommand","ngc registry resource download-version nvidia/dle/transformerxl_tf1:20.06.0","helmFetchCommand","viewAllVersionsPath","/orgs/nvidia/dle/resources/transformerxl_tf1/-/version-history","accessCheckResult",{"_208":89},"hasRelatedCollections","context",{"_204":205,"_206":207},1789216206688,[],[],[197,198,199,200,201,202,203],{"_48":63,"_50":73,"_52":65,"_54":66,"_56":70,"_58":59},{"_48":63,"_50":71,"_52":145,"_54":-5,"_56":-5,"_58":59},{"_48":63,"_50":69,"_52":65,"_54":66,"_56":70,"_58":59},{"_48":63,"_50":68,"_52":65,"_54":66,"_56":67,"_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",1663857841055],"totalSizeInBytes","malwareScanStatus","releaseNotes","### Changelog\n\nJune 2020\n * upgrade the TensorFlow container to 20.06\n * update performance tables to include A100 results\n\n* April 2020\n * Initial release\n * Support for FP32 and mixed precision training on NVIDIA\n DGX-1, NVIDIA DGX-2, and inference on NVIDIA Tesla V100 16G\n and NVIDIA T4\n\n### Known issues\n\nThere are no known issues with this model.","hasReleaseNotes","setup","The following section lists the requirements that you need to meet in order to\nstart training the Transformer-XL model.\n\n### Requirements\n\nThis repository contains `Dockerfile` which extends the TensorFlow NGC container\nand encapsulates some dependencies. Aside from these dependencies, ensure you\nhave the following components:\n\n* [NVIDIA Docker](https://github.com/NVIDIA/nvidia-docker)\n* [TensorFlow 20.06-tf1-py3](https://ngc.nvidia.com/catalog/containers/nvidia:tensorflow) NGC container\n* GPU-based architecture:\n * [NVIDIA Volta](https://www.nvidia.com/en-us/data-center/volta-gpu-architecture/)\n * [NVIDIA Turing](https://www.nvidia.com/en-us/geforce/turing/)\n * [NVIDIA Ampere architecture](https://www.nvidia.com/en-us/data-center/nvidia-ampere-gpu-architecture/)\n\nFor more information about how to get started with NGC containers, see the\nfollowing sections from the NVIDIA GPU Cloud Documentation and the Deep\nLearning DGX 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 TensorFlow](https://docs.nvidia.com/deeplearning/frameworks/tensorflow-release-notes/running.html#running)\n\nFor those unable to use the TensorFlow NGC container, to set up the required environment or create your own container,\nsee the versioned [NVIDIA Container Support\nMatrix](https://docs.nvidia.com/deeplearning/frameworks/support-matrix/index.html).","hasSetup","quickStartGuide","To train your model using mixed precision with Tensor Cores or using FP32,\nperform the following steps using the default parameters of the Transformer-XL\nbase model on the\n[WikiText-103](https://blog.einstein.ai/the-wikitext-long-term-dependency-language-modeling-dataset/)\ndataset. \n\nFor the specifics concerning training\nand inference, see the [Advanced](#advanced) section.\n\n1. Clone the repository.\n\n```\ngit clone https://github.com/NVIDIA/DeepLearningExamples\ncd DeepLearningExamples/TensorFlow/LanguageModeling/Transformer-XL\n```\n\n2. Download and preprocess the dataset.\n\n```\nbash getdata.sh\n```\n\n3. Build the Transformer-XL TensorFlow NGC container.\n\n```\nbash tf/scripts/docker/build.sh\n```\n\n4. Start an interactive session in the NGC container to run training/inference.\n\n```\nbash tf/scripts/docker/interactive.sh\n```\n\n5. Create tfrecords before your first training/evaluation for a given batch size per GPU.\nUse same --batch_chunk and --training_batch_size flags as in the training.\n\nFor training on DGX-A100 without gradient accumulation:\n```\nbash run_wt103_base.sh train_data\n```\n\nFor training on DGX-1 with gradient accumulation in 2 steps:\n```\nbash run_wt103_base.sh train_data --batch_chunk 2\n```\n\nFor single GPU training with gradient accumulation in 16 steps:\n```\nbash run_wt103_base.sh train_data --batch_chunk 16\n```\n\nFor evaluation:\n```\nbash run_wt103_base.sh test_data\n```\n\n6. Start training.\n\nTo start TF32 training on 8 GPUs on DGX-A100, run:\n\n```\nbash run_wt103_base.sh train 8\n```\n\nTo start mixed precision training on 8 GPUs on DGX-1, run:\n\n```\nbash run_wt103_base.sh train 8 --amp --batch_chunk 2\n```\n\nTo start FP32 training on single GPU, run:\n\n```\nbash run_wt103_base.sh train 1 --batch_chunk 16\n```\n\nTo start mixed precision training on 16 GPUs on DGX-2, run:\n\n```\nbash run_wt103_base.sh train 16 --amp\n```\n\nTo start FP32 training on 16 GPUs on DGX-2, run:\n\n```\nbash run_wt103_base.sh train 16\n```\n\nFor more information on the available options, and for an explanation of what\nhappens at the end of training, refer to the [Training\nprocess](#training-process) section.\n\n7. Start evaluation.\n\nTo start mixed precision inference on the test set, run:\n\n```\nbash run_wt103_base.sh eval [--amp]\n```\n\nThe `--amp` flag is optional, however, if it's set, then the script\nlaunches mixed precision inference with Tensor Cores. If the flag is not\npresent, then the script launches FP32 inference.\nBy default, the script is loading the checkpoint from\n`LM-TFM/model.ckpt`, which contains the model corresponding to the\nlast checkpoint from the previous training run. The path to the\ncheckpoint can be customized by setting the `--model_dir` flag.\n\nFor more information on the available options, refer to the [Inference\nprocess](#inference-process) section.","hasQuickStartGuide","advanced","The following sections provide greater details of the dataset, running training\nand inference, and the training results.\n\n### Scripts and sample code\n\n* `Dockerfile`: a container with the basic set of dependencies to run\n Transformer-XL\n\nIn the `tf` directory, the most important files are:\n\n* `data_utils.py`: data loading utilities\n* `exp_utils.py`: utility functions for running training and benchmarking\n* `lamb.py`: implementation of [LAMB](https://arxiv.org/abs/1904.00962)\n optimizer\n* `main.py`: serves as the entry point to launch the training and inference\n* `model.py`: implementation of the Transformer-XL model\n* `vocabulary.py`: implementation of word-level vocabulary\n\n### Parameters\n\nThe complete list of available parameters for the `tf/main.py` script contains:\n\n```\n --batch_chunk: Number of accumulation steps.\n (default: '1')\n (an integer)\n --clamp_len: Clamp length\n (default: '-1')\n (an integer)\n --clip: Gradient clipping value.\n (default: '0.25')\n (a number)\n --corpus_info_path: Path to corpus-info.json file.\n (default: '')\n --d_embed: Dimension of the embeddings.\n (default: '512')\n (an integer)\n --d_head: Dimension of each attention head.\n (default: '64')\n (an integer)\n --d_inner: Dimension of inner hidden size in positionwise feed-forward.\n (default: '2048')\n (an integer)\n --d_model: Dimension of the model.\n (default: '512')\n (an integer)\n --data_dir: Path to tf-records directory.\n (default: '')\n --div_val: Divide the embedding size by this val for each bin\n (default: '1')\n (an integer)\n --[no]do_eval: Whether to run eval on the dev set.\n (default: 'false')\n --[no]do_train: Whether to run training.\n (default: 'true')\n --dropatt: Attention dropout rate.\n (default: '0.0')\n (a number)\n --dropout: Dropout rate.\n (default: '0.1')\n (a number)\n --eval_batch_size: Size of valid batch.\n (default: '16')\n (an integer)\n --eval_ckpt_path: Checkpoint path for do_test evaluation.If set, model_dir will be ignored.If unset, will use the latest ckpt in model_dir.\n --eval_split: Which data split to evaluate.\n (default: 'valid')\n --amp: Whether to enable AMP ops.\n (default: 'false')\n --init: : Initialization method.\n (default: 'normal')\n --init_range: Initialization std when init is uniform.\n (default: '0.1')\n (a number)\n --init_std: Initialization std when init is normal.\n (default: '0.02')\n (a number)\n --learning_rate: Maximum learning rate.\n (default: '0.01')\n (a number)\n --log_interval: Number of iterations per repeat loop.\n (default: '100')\n (an integer)\n --max_eval_batch: Set -1 to turn off. Only used in test mode.\n (default: '-1')\n (an integer)\n --mem_len: Number of steps to cache\n (default: '192')\n (an integer)\n --min_lr_ratio: Minimum ratio learning rate.\n (default: '0.1')\n (a number)\n --model_dir: Estimator model_dir.\n (default: 'LM-TFM')\n --n_head: Number of attention heads.\n (default: '8')\n (an integer)\n --n_layer: Number of layers.\n (default: '16')\n (an integer)\n --num_core_per_host: Number of cores per host\n (default: '8')\n (an integer)\n --percentiles: percentiles for latency confidence intervals\n (default: '90,95,99')\n (a comma separated list)\n --proj_init_std: Initialization std for embedding projection.\n (default: '0.01')\n (a number)\n --[no]proj_same_dim: Project the bin with the same dimension.\n (default: 'true')\n --[no]proj_share_all_but_first: True to share all but first projs, False not to share.\n (default: 'false')\n --record_info_dir: Path to local directory containing filenames.txt.\n (default: '')\n --[no]same_length: Same length attention\n (default: 'false')\n --save_steps: number of steps for model checkpointing.\n (default: '5000')\n (an integer)\n --tgt_len: Number of steps to predict\n (default: '192')\n (an integer)\n --[no]tie_weight: Tie embedding and softmax weight.\n (default: 'true')\n --train_batch_size: Size of train batch.\n (default: '256')\n (an integer)\n --train_steps: Total number of training steps.\n (default: '40000')\n (an integer)\n --[no]untie_r: untie r_w_bias and r_r_bias\n (default: 'false')\n --warmup_steps: Number of steps for linear lr warmup.\n (default: '1000')\n (an integer)\n```\n\n### Command-line options\n\nTo see the full list of available options and their descriptions, use the `--help` command-line option.\nFor example:\n\n```\npython3 main.py --help\n```\n\n### Getting the data\n\nThe Transformer-XL model was trained on the\n[WikiText-103](https://blog.einstein.ai/the-wikitext-long-term-dependency-language-modeling-dataset/)\ndataset. The WikiText-103 dataset is a collection of over 100 million tokens\nextracted from the set of verified\n[Good](https://en.wikipedia.org/wiki/Wikipedia:Good_articles) and\n[Featured](https://en.wikipedia.org/wiki/Wikipedia:Featured_articles) articles\non Wikipedia.\n\nThis repository contains the `getdata.sh` download script which\nautomatically downloads and extracts the training, validation and test\ndatasets. By default, data is downloaded to the `data` directory.\n\nIn order to test with other datasets, the script needs to be customized\naccordingly.\n\n#### Dataset guidelines\n\nThe WikiText-103 dataset was already pre-tokenized with word-level tokens. The\ndataset features a large vocabulary of 267,735 tokens and retains the original\ncase, punctuation and numbers.\n\nThe `getdata.sh` script downloads the data, extracts the archive and renames\nthe training, validation, and test set to `train.txt`, `valid.txt`, `test.txt`\nrespectively.\n\n#### Multi-dataset\n\nUsing other datasets requires changes in the `tf/data_utils.py` file:\n* the name of the new dataset should be added to the `dataset` flag\n* the support for the new dataset needs to be added to the `Corpus` class:\n names of files containing training, validation and test data, options for\n the tokenizer, dataset iterator and desired values of cutoffs for adaptive softmax\n\nThe current codebase supports training with word-level vocabulary\n(automatically generated based on the provided dataset)\n\nAdditionally, using other datasets may require changes in some hyperparameters\n(for example, batch size, learning rate, number of training steps,\nand the configuration of learning rate scheduler). \n\n### Training process\n\nThe default training configuration can be launched by running the\n`run_wt103_base.sh` script with the first argument\nset to `train`. By default, the training results are saved to `tf/LM-TFM` directory,\nand map to your container's `/workspace/transformer-x/tf/LM-TFM` directory;\nthis can be customized by setting the `--model_dir` parameter.\n\nThe training script launches a single-node data-parallel training with a fixed\nglobal batch size of 256, optionally with gradient accumulation to allow\ntraining on configurations with less than 16 GPUs.\n\n**Command-line**\n\nYou can launch training of the Transformer-XL base model on the\nWikiText-103 dataset with the word-based vocabulary and adaptive softmax using\n`<#GPUs>` GPUs. For example:\n\n```\nbash run_wt103_base.sh train <#GPUs> [--amp] [--batch_chunk CHUNK]\n```\n\nThe `--amp` flag is optional, however, if it's set, then the script\nlaunches mixed precision training with Tensor Cores; if the flag is not\npresent, then the script launches FP32 training.\n\nThe `--batch_chunk CHUNK` parameter controls gradient accumulation. With\ngradient accumulation, the batch size is split into `CHUNK` chunks of equal\nsize, the training script executes the forward and backward pass using each\nchunk and then executes the optimizer using accumulated gradients.\n\nThe `--train batch_size GBS` parameter can be used to change batch size.\nEquation below shows the relation between parameters:\n```\nglobal batch size (--train_batch_size) = <#GPU> * gradient accumulation steps (--batch_chunk) * local batch size\n```\n\n**Examples**\n\nYou can launch mixed precision training of the Transformer-XL base model on the\nWikiText-103 dataset using 16 GPUs. For example:\n\n```\nbash run_wt103_base.sh train 16 --amp --batch_chunk 1\n```\n\nThe batch size per GPU is equal to the default global batch size of 256\ndivided by the product of the number of GPUs times the number of chunks. In this\ncase, batch size per GPU is equal to `256 / (16 * 1) = 16`.\n\nYou can launch FP32 training using 8 GPUs; the batch size per GPU is equal to 16\n(`--batch_chunk` was set to `2` because a local batch size of 32 runs out\nof memory on a DGX-1 with Tesla V100 16G in FP32 training). For example:\n\n```\nbash run_wt103_base.sh train 8 --batch_chunk 2\n```\n\nA summary of the training progress is printed after every 100 training\niterations; this can be customized by setting the `--log_interval` parameter.\nThe summary is printed in the following format:\n\n```\nstep 1300 | lr 0.009998686 | loss 5.09 | pplx 162.70, bpc 7.3461, tok/s 138037\n```\n\nwhich contains information about a current training\nstep, current learning rate, current training loss,\ntraining [perplexity](https://en.wikipedia.org/wiki/Perplexity#Perplexity_per_word),\nbits per character and throughput in tokens per second.\n\nThe script saves one checkpoint: `model.ckpt` which contains the last saved model.\nBy default, model saving is executed every\n5000 training steps, this can be customized by setting the `--save_steps`\nparameter.\n\nEvaluation (inference) benefits from longer attention sequences, therefore to\nreproduce perplexity values reported in the [Transformer-XL\npaper](https://arxiv.org/abs/1901.02860), it's necessary to run the final\nevaluation with a dedicated inference script. Refer to the [Inference\nprocess](#inference-process) section for more details.\n\n### Inference process\n\nInference can be run by launching the `run_wt103_base.sh` script\nwith the first argument set to `eval`. Running\ninference requires a pre-trained model checkpoint.\n\nThe script supports only single-GPU inference.\n\n**Command-line**\n\nYou can launch inference of the Transformer-XL base model on the\nWikiText-103 dataset with the word-based vocabulary and adaptive softmax.\n\nFor example:\n\n```\nbash run_wt103_base.sh eval --model_dir [--amp]\n```\n\nThe `--amp` flag is optional, however, if it's specified, then the script\nlaunches inference with Tensor Cores; if the flag is not present, then the\nscript launches FP32 inference.\n\n**Examples**\n\nTo launch mixed precision inference on a single GPU using a checkpoint\nloaded from `LM-TFM/model.ckpt*`, run:\n\n```\nbash run_wt103_base.sh eval --model_dir LM-TFM --amp\n```\n\nTo launch FP32 inference on a single GPU using a checkpoint loaded\nfrom `LM-TFM/model.ckpt*`, run:\n\n```\nbash run_wt103_base.sh eval --model_dir LM-TFM\n```\n\nAfter the execution, the script prints a summary in the following format:\n\n```\nI0109 13:02:31.304439 139903273469760 main.py:440] Evaluating with: math fp16\nINFO:tensorflow:| loss 3.15 | pplx 23.32, bpc 4.5432, tok/s 9946, ms/batch 102.84\n```\n\nwhich contains information about loss, perplexity and execution performance on the test dataset.","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\n\nThe following section shows how to run benchmarks measuring the model\nperformance in training and inference modes.\n\n#### Training performance benchmark\n\nTo benchmark the training performance on a specific global batch size ``,\nwith a specific number of GPUs `<#GPUs>` for a specific number of training\niterations `` run:\n\nFor the base model:\n\n```\nbash run_wt103_base.sh train <#GPUs> --train_batch_size --train_steps --log_interval 1 [--amp] [--batch_chunk CHUNK]\n```\n\nIt's recommended to launch at least 1500 training steps to get a reliable\nestimate of training performance. For more information about the available\noptions, refer to the [Training process](#training-process) section.\n\nThe training script prints information in the following format:\n\n```\n(...)\n[1,0]:INFO:tensorflow:step 99 | lr 0.000990000 | loss 9.22 | pplx 10069.60, bpc 13.2977, tok/s 136092\n[1,0]:I0109 12:18:41.333325 140403024426816 main.py:333] step 99 | lr 0.000990000 | loss 9.22 | pplx 10069.60,\nbpc 13.2977, tok/s 136092\n[1,0]:INFO:tensorflow:step 100 | lr 0.001000000 | loss 9.21 | pplx 9981.87, bpc 13.2851, tok/s 135309\n[1,0]:I0109 12:18:41.696926 140403024426816 main.py:333] step 100 | lr 0.001000000 | loss 9.21 | pplx 9981.87,\nbpc 13.2851, tok/s 135309\n(...)\n[1,0]:INFO:tensorflow:Training throughput: 135959 tok/s\n```\n\nThe last two lines contain information on the\naverage training throughput measured in tokens per second.\n\n#### Inference performance benchmark\n\nThe inference performance and accuracy benchmarks require a checkpoint from a\ntrained model.\n\nTo benchmark the inference performance on a specific global batch size ``, run:\n\n```\nbash run_wt103_base.sh eval --model_dir --eval_batch_size [--amp]\n```\n\nThe inference script prints information in the following format:\n\n```\nI0109 13:02:31.304439 139903273469760 main.py:440] Evaluating with: math fp16\nINFO:tensorflow:| loss 3.15 | pplx 23.32, bpc 4.5432, tok/s 9946, ms/batch 102.84\n```\n\nThe output contains information on the achieved test loss and test perplexity,\naverage inference throughput (measured in tokens per second), average inference\nlatency (measured in milliseconds).\n\n### Results\n\nThe following sections provide details on how we achieved our performance and\naccuracy in training and inference.\n\n#### Training accuracy results\n\n##### Training accuracy: NVIDIA DGX A100 (8x A100 40GB)\n\n###### Base model\nOur results were obtained by running the `tf/run_wt103_base.sh`\ntraining script in the tensorflow:20.06-tf1-py3 NGC container on NVIDIA DGX A100\nwith8x A100 40GB GPUs.\n\n|**GPUs**|**Batch Size / GPU**|**Accuracy - TF32 (perplexity)**|**Accuracy - Mixed precision (perplexity)**|**Time to Train - TF32 (minutes)**|**Time to Train - Mixed precision (minutes)**|**Time to Train Speedup (TF32 to Mixed precision)**|\n|-------:|-------------------:|-------------------------------:|------------------------------------------:|---------------------------------:|--------------------------------------------:|--------------------------------------------------:|\n| 1 | 16 | 23.53 | 23.50 | 960 | 880 | 1.09 |\n| 8 | 16 | 23.45 | 23.48 | 150 | 142 | 1.06 |\n\n##### Training accuracy: NVIDIA DGX-1 (8x V100 16GB)\n\n###### Base model\nOur results were obtained by running the `tf/run_wt103_base.sh`\ntraining script in the tensorflow:20.06-tf1-py3 NGC container on NVIDIA DGX-1\nwith 8x V100 16G GPUs.\n\n|**GPUs**|**Batch Size / GPU**|**Accuracy - FP32 (perplexity)**|**Accuracy - Mixed precision (perplexity)**|**Time to Train - FP32 (minutes)**|**Time to Train - Mixed precision (minutes)**|**Time to Train Speedup (FP32 to Mixed precision)**|\n|-------:|-------------------:|-------------------------------:|------------------------------------------:|---------------------------------:|--------------------------------------------:|--------------------------------------------------:|\n| 1 | 16 | 23.64 | 23.58 | 2949 | 2021 | 1.46 |\n| 8 | 16 | 23.35 | 23.34 | 459 | 343 | 1.34 |\n\n##### Training accuracy: NVIDIA DGX-2 (16x V100 32GB)\n\n###### Base model\n\nOur results were obtained by running the `tf/run_wt103_base.sh`\ntraining script in the tensorflow:20.06-tf1-py3 NGC container on NVIDIA DGX-2\nwith 16x V100 32G GPUs.\n\n|**GPUs**|**Batch Size / GPU**|**Accuracy - FP32 (perplexity)**|**Accuracy - Mixed precision (perplexity)**|**Time to Train - FP32 (minutes)**|**Time to Train - Mixed precision (minutes)**|**Time to Train Speedup (FP32 to Mixed precision)**|\n|-------:|-------------------:|-------------------------------:|------------------------------------------:|---------------------------------:|--------------------------------------------:|--------------------------------------------------:|\n| 16 | 16 | 23.39 | 23.37 | 202 | 161 | 1.25 |\n| 8 | 32 | 23.33 | 23.40 | 330 | 227 | 1.46 |\n\n##### Training loss plot\n\n###### Base model\n\n![TrainingLossBase](https://raw.githubusercontent.com/NVIDIA/DeepLearningExamples/master/TensorFlow/LanguageModeling/Transformer-XL//tf/img/training_loss_base.png)\n\n\n##### Training stability test\n\n###### Base model\nThe Transformer-XL base model was trained for 40,000 training steps, starting\nfrom 20 different initial random seeds. The training was performed in the tensorflow:20.06-tf1-py3 NGC container on\nNVIDIA DGX-1 with 8x V100 16G GPUs.\nAfter training, the models were evaluated on the test dataset. The following\ntable summarizes the final perplexity on the test set.\n\n|**Average perplexity**|**Standard deviation**|**Minimum**|**Maximum**|**Median**|\n|---------------------:|---------------------:|----------:|----------:|---------:|\n| 23.38 | 0.0879 | 23.24 | 23.58 | 23.39 |\n\n#### Training performance results\n\n##### Training performance: NVIDIA DGX A100 (8x A100 40GB)\n\nOur results were obtained by running the `tf/run_wt103_base.sh`\ntraining script in the tensorflow:20.06-tf1-py3 NGC container on NVIDIA DGX A100\nwith 8x A100 40GB GPUs. Performance numbers (in tokens per second) were averaged over 2000\ntraining iterations.\n\n|**GPUs**|**Batch Size / GPU**|**Throughput - TF32 (tok/s)**|**Throughput - Mixed precision (tok/s)**|**Throughput speedup (TF32 to Mixed precision)**|**Weak Scaling - TF32**|**Weak Scaling - Mixed precision**|\n|-------:|-------------------:|----------------------------:|---------------------------------------:|-----------------------------------------------:|----------------------:|---------------------------------:|\n| 1 | 16 | 25,127 | 26,130 | 1.040 | 1.000 | 1.000 |\n| 1 | 32 | 30,958 | 33,117 | 1.070 | 1.000 | 1.000 |\n| 1 | 64 | 34,244 | 36,455 | 1.065 | 1.000 | 1.000 |\n| 8 | 16 | 157,538 | 155,656 | 0.988 | 6.270 | 5.957 |\n| 8 | 32 | 224,474 | 227,502 | 1.013 | 7.251 | 6.870 |\n\nTo achieve these same results, follow the steps in the [Quick Start Guide](#quick-start-guide).\n\n##### Training performance: NVIDIA DGX-1 (8x V100 16GB)\n\n###### Base model\n\nOur results were obtained by running the `tf/run_wt103_base.sh`\ntraining script in the tensorflow:20.06-tf1-py3 NGC container on NVIDIA DGX-1 with 8x\nV100 16G GPUs. Performance numbers (in tokens per second) were averaged over 2000\ntraining iterations.\n\n|**GPUs**|**Batch Size / GPU**|**Throughput - FP32 (tok/s)**|**Throughput - Mixed precision (tok/s)**|**Throughput speedup (FP32 to Mixed precision)**|**Weak Scaling - FP32**|**Weak Scaling - Mixed precision**|\n|-------:|-------------------:|----------------------------:|---------------------------------------:|-----------------------------------------------:|----------------------:|---------------------------------:|\n| 1 | 16 | 9,104 | 13,004 | 1.428 | 1.000 | 1.000 |\n| 2 | 16 | 18,169 | 23,856 | 1.313 | 1.996 | 1.835 |\n| 4 | 16 | 38,876 | 50,310 | 1.294 | 4.270 | 3.869 |\n| 8 | 16 | 78,626 | 101,954 | 1.297 | 8.636 | 7.840 |\n\nTo achieve these same results, follow the steps in the [Quick Start Guide](#quick-start-guide).\n\n##### Training performance: NVIDIA DGX-2 (16x V100 32GB)\n\n###### Base model\n\nOur results were obtained by running the `tf/run_wt103_base.sh` training\nscript in the tensorflow:20.06-tf1-py3 NGC container on NVIDIA DGX-2 with 16x V100 32G\nGPUs. Performance numbers (in tokens per second) were averaged over 2000\ntraining iterations.\n\n|**GPUs**|**Batch Size / GPU**|**Throughput - FP32 (tok/s)**|**Throughput - Mixed precision (tok/s)**|**Throughput speedup (FP32 to Mixed precision)**|**Weak Scaling - FP32**|**Weak Scaling - Mixed precision**|\n|-------:|-------------------:|----------------------------:|---------------------------------------:|-----------------------------------------------:|----------------------:|---------------------------------:|\n| 1 | 16 | 9,891 | 13,791 | 1.394 | 1.000 | 1.000 |\n| 2 | 16 | 21,550 | 28,306 | 1.314 | 2.179 | 2.052 |\n| 4 | 16 | 42,616 | 55,430 | 1.301 | 4.309 | 4.019 |\n| 8 | 16 | 83,932 | 107,999 | 1.287 | 8.486 | 7.831 |\n| 16 | 16 | 164,675 | 206,906 | 1.256 | 16.649 | 15.003 |\n\nTo achieve these same results, follow the steps in the [Quick Start Guide](#quick-start-guide).\n\n#### Inference performance results\n\n##### Inference performance: NVIDIA DGX A100 (1x A100 40GB)\n\n###### Base model\n\nOur results were obtained by running the\n`tf/scripts/inference_benchmark.sh` inferencing benchmarking script in the\ntensorflow:20.06-tf1-py3 NGC container on NVIDIA DGX A100 (1x A100 40GB) GPU.\n\nThe command to launch the inference performance benchmark is provided in the\n[Inference performance benchmark](#inference-performance-benchmark) section.\n\n**FP16**\n\n|**Batch size**|**Sequence length**|**Memory length**|**Throughput Avg (tok/s)**|**Latency Avg (ms)**|**Latency 90% (ms)**|**Latency 95% (ms)**|**Latency 99% (ms)**|\n|-------------:|------------------:|----------------:|-------------------------:|-------------------:|-------------------:|-------------------:|-------------------:|\n| 1 | 64 | 640 | 2592.8 | 24.71 | 25.72 | 26.12 | 26.68 |\n| 2 | 64 | 640 | 5060.4 | 25.32 | 26.58 | 26.93 | 27.71 |\n| 4 | 64 | 640 | 8910.2 | 28.73 | 29.74 | 30.06 | 30.58 |\n| 8 | 64 | 640 | 13844.1 | 36.96 | 37.62 | 37.80 | 38.34 |\n| 16 | 64 | 640 | 18313.1 | 55.92 | 56.46 | 56.69 | 57.39 |\n| 32 | 64 | 640 | 21854.7 | 93.63 | 94.37 | 94.74 | 94.92 |\n\n**TF32**\n\n|**Batch size**|**Sequence length**|**Memory length**|**Throughput Avg (tok/s)**|**Latency Avg (ms)**|**Latency 90% (ms)**|**Latency 95% (ms)**|**Latency 99% (ms)**|\n|-------------:|------------------:|----------------:|-------------------------:|-------------------:|-------------------:|-------------------:|-------------------:|\n| 1 | 64 | 640 | 2587.6 | 24.75 | 25.63 | 25.92 | 26.44 |\n| 2 | 64 | 640 | 5177.8 | 24.73 | 25.71 | 26.03 | 26.56 |\n| 4 | 64 | 640 | 9113.6 | 28.09 | 29.40 | 29.71 | 30.07 |\n| 8 | 64 | 640 | 13371.7 | 38.27 | 38.95 | 39.34 | 40.07 |\n| 16 | 64 | 640 | 16971.0 | 60.29 | 60.88 | 61.13 | 61.73 |\n| 32 | 64 | 640 | 19434.5 | 105.29 | 106.00 | 106.19 | 106.79 |\n\nTo achieve these same results, follow the steps in the [Quick Start Guide](#quick-start-guide).\n\n##### Inference performance: NVIDIA DGX-1 (1x V100 16GB)\n\n###### Base model\n\nOur results were obtained by running the\n`tf/scripts/inference_benchmark.sh` inferencing benchmarking script in the\ntensorflow:20.06-tf1-py3 NGC container on NVIDIA DGX-1 with 1x V100 16G GPU.\n\nThe command to launch the inference performance benchmark is provided in the\n[Inference performance benchmark](#inference-performance-benchmark) section.\n\n**FP16**\n\n|**Batch size**|**Sequence length**|**Memory length**|**Throughput Avg (tok/s)**|**Latency Avg (ms)**|**Latency 90% (ms)**|**Latency 95% (ms)**|**Latency 99% (ms)**|\n|-------------:|------------------:|----------------:|-------------------------:|-------------------:|-------------------:|-------------------:|-------------------:|\n| 1 | 64 | 640 | 1823.8 | 35.17 | 37.27 | 38.22 | 41.28 |\n| 2 | 64 | 640 | 3337.6 | 38.46 | 41.09 | 41.94 | 43.91 |\n| 4 | 64 | 640 | 5354.0 | 47.83 | 49.74 | 50.54 | 53.08 |\n| 8 | 64 | 640 | 7779.7 | 65.79 | 67.71 | 68.37 | 69.71 |\n| 16 | 64 | 640 | 9796.5 | 104.46 | 107.22 | 108.07 | 108.69 |\n| 32 | 64 | 640 | 11215.5 | 182.45 | 184.11 | 184.49 | 186.92 |\n\n**FP32**\n\n|**Batch size**|**Sequence length**|**Memory length**|**Throughput Avg (tok/s)**|**Latency Avg (ms)**|**Latency 90% (ms)**|**Latency 95% (ms)**|**Latency 99% (ms)**|\n|-------------:|------------------:|----------------:|-------------------------:|-------------------:|-------------------:|-------------------:|-------------------:|\n| 1 | 64 | 640 | 1912.7 | 33.56 | 35.98 | 36.84 | 38.96 |\n| 2 | 64 | 640 | 3497.0 | 36.66 | 39.07 | 39.85 | 41.28 |\n| 4 | 64 | 640 | 4732.9 | 54.10 | 56.32 | 57.10 | 58.14 |\n| 8 | 64 | 640 | 6303.7 | 81.19 | 83.32 | 84.02 | 88.12 |\n| 16 | 64 | 640 | 7676.3 | 133.29 | 134.84 | 135.33 | 136.70 |\n| 32 | 64 | 640 | 8555.6 | 239.15 | 240.02 | 240.20 | 240.48 |\n\nTo achieve these same results, follow the steps in the [Quick Start Guide](#quick-start-guide).\n\n##### Inference performance: NVIDIA T4\n\n###### Base model\n\nOur results were obtained by running the\n`tf/scripts/inference_benchmark.sh` inferencing benchmarking script in the\ntensorflow:20.06-tf1-py3 NGC container on NVIDIA T4.\n\nThe command to launch the inference performance benchmark is provided in the\n[Inference performance benchmark](#inference-performance-benchmark) section.\n\n**FP16**\n\n|**Batch size**|**Sequence length**|**Memory length**|**Throughput Avg (tok/s)**|**Latency Avg (ms)**|**Latency 90% (ms)**|**Latency 95% (ms)**|**Latency 99% (ms)**|\n|-------------:|------------------:|----------------:|-------------------------:|-------------------:|-------------------:|-------------------:|-------------------:|\n| 1 | 64 | 640 | 1228.6 | 52.21 | 55.14 | 56.32 | 59.77 |\n| 2 | 64 | 640 | 2108.5 | 60.78 | 63.62 | 64.68 | 67.07 |\n| 4 | 64 | 640 | 3376.7 | 75.83 | 78.77 | 79.63 | 82.80 |\n| 8 | 64 | 640 | 4666.3 | 109.69 | 112.58 | 113.88 | 117.35 |\n| 16 | 64 | 640 | 5557.0 | 184.14 | 186.51 | 187.20 | 189.64 |\n| 32 | 64 | 640 | 6174.3 | 331.41 | 333.67 | 334.94 | 336.90 |\n\n**FP32**\n\n|**Batch size**|**Sequence length**|**Memory length**|**Throughput Avg (tok/s)**|**Latency Avg (ms)**|**Latency 90% (ms)**|**Latency 95% (ms)**|**Latency 99% (ms)**|\n|-------------:|------------------:|----------------:|-------------------------:|-------------------:|-------------------:|-------------------:|-------------------:|\n| 1 | 64 | 640 | 1029.7 | 62.28 | 65.82 | 66.93 | 70.22 |\n| 2 | 64 | 640 | 1667.7 | 76.81 | 79.80 | 80.71 | 84.35 |\n| 4 | 64 | 640 | 2302.3 | 111.13 | 113.75 | 114.85 | 118.57 |\n| 8 | 64 | 640 | 2756.9 | 185.58 | 188.16 | 189.38 | 192.68 |\n| 16 | 64 | 640 | 3188.8 | 320.86 | 324.24 | 325.63 | 327.76 |\n| 32 | 64 | 640 | 3439.1 | 594.96 | 599.13 | 599.89 | 602.59 |\n\nTo achieve these same results, follow the steps in the [Quick Start Guide](#quick-start-guide).","hasPerformance",[],{"_84":181,"_111":233,"_112":-5,"_120":34,"_119":34,"_121":232,"_216":89},{"_84":28,"_111":28,"_112":-5,"_120":34,"_119":34,"_121":232,"_216":89},[],"Latest (20.06.0)","nvidia/dle/transformerxl_tf1","key","nvidia/dle/resources/transformerxl_tf1","resourceId","transformerxl_tf1","shortDescription","Transformer-XL is a transformer-based language model with a segment-level recurrence and a novel relative positional encoding.","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/transformerxl-tf1)\n\n\nThis repository provides an implementation of the Transformer-XL model in\n[TensorFlow](https://www.tensorflow.org) from the paper [Transformer-XL: Attentive\nLanguage Models Beyond a Fixed-Length\nContext](https://arxiv.org/abs/1901.02860). Transformer-XL is a\ntransformer-based language model with a segment-level recurrence and a novel\nrelative positional encoding. Enhancements introduced in Transformer-XL help\ncapture better long-term dependencies by attending to tokens from multiple\nprevious segments.\n\nOur implementation is based on the\n[codebase](https://github.com/kimiyoung/transformer-xl) published by the\nauthors of the Transformer-XL paper.\nOur implementation uses a modified model architecture. Our\nmodifications were made to achieve better hardware utilization and to take\nadvantage of Tensor Cores. Similar modifications were also proposed in an\nimplementation available from\n[github.com/cybertronai/transformer-xl](https://github.com/cybertronai/transformer-xl).\nRefer to the [Model architecture](#model-architecture) section for more\ndetails.\n\nThis model is trained with mixed precision using Tensor Cores on Volta, Turing,\nand the NVIDIA Ampere GPU architectures. Therefore, researchers can get\nresults up to 1.5x faster than training without Tensor Cores, while\nexperiencing the benefits of mixed precision training. This model is tested\nagainst each NGC monthly container release to ensure consistent accuracy and\nperformance over time.\n\n### Model architecture\n\nThe Transformer-XL \"base\" model for WikiText-103 dataset available in this\nrepository was modified to use the following hyperparameter values:\n\n|**Hyperparameter**|**Description**|**Original setting for the base model**|**Our modification to the base model**|\n|------------------|---------------|--------------------------------------:|--------------------------------------:|\n| `d_model` | hidden size | 410 | 512 |\n| `n_head` | number of attention heads | 10 | 8 |\n| `d_head` | size of each attention head | 41 | 64 |\n| `d_inner` | hidden size in fully-connected layers | 2100 | 2048 |\n| `tgt_len` | number of tokens to predict during training | 150 | 192 |\n| `mem_len` | number of tokens cached from previous iterations during training | 150 | 192 |\n\nChanges described above were made to align certain hyperparameters with powers\nof two, with this modification, the model is able to achieve better hardware\nutilization, and therefore higher training throughput.\n\nThe following table lists the hyperparameters for the base\nTransformer-XL model for WikiText-103 dataset available in this repository.\n\n| **Hyperparameter** | **Description** | **Base model** |\n| ------------------ | ---------------------------------------------------------------- | -------------: |\n| `n_layer` | number of layers | 16 |\n| `d_model` | hidden size | 512 |\n| `n_head` | number of attention heads | 8 |\n| `d_head` | size of each attention head | 64 |\n| `d_inner` | inner hidden size in fully-connected layers | 2048 |\n| `dropout` | dropout | 0.1 |\n| `dropatt` | dropout after softmax in the attention | 0.0 |\n| `lr` | base learning rate | 0.01 |\n| `min_lr_ratio` | minimum ratio learning rate (for cosine decay) | 0.1 |\n| `max_step` | number of training steps | 40,000 |\n| `warmup_step` | number of learning rate warmup steps | 1,000 |\n| `batch_size` | training batch size | 256 |\n| `tgt_len` | number of tokens to predict during training | 192 |\n| `mem_len` | number of tokens cached from previous iterations during training | 192 |\n\nThe Transformer-XL model addresses the limitations of vanilla transformer-based\nlanguage models, which are only able to use relatively short context, bounded\nby the segment length. The Transformer-XL introduces a recurrence mechanism,\nwhich is able to use a cached hidden state from previous segments. During\ntraining, the context consists of a concatenation of the current segment's hidden\nstate and cached states from previous iterations. Gradients are backpropagated\nonly through the current segment, although the model is able to take advantage\nof the extra information stored in the cache and therefore is able to model\nlong-term dependencies.\n\nAn illustration of the recurrence mechanism taken from the [Transformer-XL\npaper](https://arxiv.org/abs/1901.02860) is shown below.\n![model](https://raw.githubusercontent.com/NVIDIA/DeepLearningExamples/master/TensorFlow/LanguageModeling/Transformer-XL//tf/img/model.png)\n\n\n### Default configuration\n\nThe following features were implemented in this model:\n\n* general\n * single-node, Horovod multi-GPU training\n * training and inference with mixed precision using Tensor Cores\n * automatic mixed precision training (AMP)\n* model\n * 16-layer base Transformer-XL model with hidden size 512, 8 attention heads,\n each head with hidden size 64\n * the model trained on\n [WikiText-103](https://blog.einstein.ai/the-wikitext-long-term-dependency-language-modeling-dataset/)\n dataset, using word-level vocabulary and\n adaptive softmax\n * embedding weights are tied with weights in the classifier\n* training\n * training with [LAMB](https://arxiv.org/abs/1904.00962) optimizer, the\n implementation of the optimizer uses [XLA](https://www.tensorflow.org/xla), which enables\n the fusion of elementwise operations and accelerates the training\n * support for training with a gradient accumulation\n * base model:\n * linear learning rate warmup for 1,000 iterations, followed by the cosine\n learning rate schedule, the initial learning rate is set to 0.0, and the final\n learning rate is set to 0.001 (min_lr_ratio * base_lr)\n * training for 40,000 steps, using a batch size of 256\n* inference\n * support for single-GPU inference\n * each token is using the same size of the context from previous time steps.\n * base model:\n * target length is set to 64, length of memory is set to 640\n * positional embeddings are clamped after 400 time steps\n\n### Feature support matrix\n\nThe following features are supported by this model:\n\n| **Feature** | **Transformer-XL** |\n|:------------|-------------------:|\n|[Automatic mixed precision (AMP)](https://nvidia.github.io/apex/amp.html) | Yes |\n|[Horovod Multi-GPU (NCCL)](https://github.com/horovod/horovod) | Yes |\n|[LAMB](https://arxiv.org/abs/1904.00962v3) | Yes |\n\n#### Features\n\n[TF-AMP](https://docs.nvidia.com/deeplearning/dgx/tensorflow-user-guide/index.html#tfamp) - a \ntool that enables Tensor Core-accelerated training. Refer to the [Enabling\nmixed precision](#enabling-mixed-precision) section for more details.\n\n[Horovod](https://github.com/horovod/horovod) - Horovod \nis a distributed training framework for TensorFlow, Keras, PyTorch, and MXNet.\nThe goal of Horovod is to make distributed deep learning fast and easy to use.\nFor more information about how to get started with Horovod, see the [Horovod:\nOfficial repository](https://github.com/horovod/horovod).\n\n[Multi-GPU training with Horovod](https://github.com/horovod/horovod/#usage) - our model \nuses Horovod to implement efficient multi-GPU training with NCCL. For details,\nsee example sources in this repository or see the [TensorFlow\ntutorial](https://github.com/horovod/horovod/#usage).\n\n[LAMB](https://arxiv.org/abs/1904.00962v3) - stands \nfor Layerwise Adaptive Moments Based optimizer, is a large batch optimization\ntechnique that helps accelerate training of deep neural networks using large\nminibatches.\n\n### Mixed precision training\n\nMixed precision is the combined use of different numerical precisions in a\ncomputational method.\n[Mixed precision](https://arxiv.org/abs/1710.03740) training offers significant\ncomputational speedup by performing operations in half-precision format while\nstoring minimal information in single-precision to retain as much information\nas possible in critical parts of the network. Since the introduction of [Tensor\nCores](https://developer.nvidia.com/tensor-cores) in Volta, and following with both the\nTuring and Ampere architectures, significant training speedups are experienced by switching to\nmixed precision -- up to 3x overall speedup on the most arithmetically intense\nmodel architectures. Using mixed precision training previously required two\nsteps:\n\n1. Porting the model to use the FP16 data type where appropriate. \n2. Adding loss scaling to preserve small gradient values.\n\nThis can now be achieved using Automatic Mixed Precision (AMP) for TensorFlow to enable the full\n[mixed precision methodology](https://docs.nvidia.com/deeplearning/sdk/mixed-precision-training/index.html#tensorflow)\nin your existing TensorFlow model code. AMP enables mixed precision training on Volta and Turing GPUs automatically.\nThe TensorFlow framework code makes all necessary model changes internally.\n\nIn TF-AMP, the computational graph is optimized to use as few casts as necessary and maximize the use of FP16,\nand the loss scaling is automatically applied inside of supported optimizers. AMP can be configured to work\nwith the existing tf.contrib loss scaling manager by disabling the AMP scaling with a single environment\nvariable to perform only the automatic mixed-precision optimization. It accomplishes this by automatically\nrewriting all computation graphs with the necessary operations to enable mixed precision training and automatic loss scaling.\n\nFor information about:\n\n* How to train using mixed precision, see the [Mixed Precision\n Training](https://arxiv.org/abs/1710.03740) paper and [Training With Mixed\n Precision](https://docs.nvidia.com/deeplearning/sdk/mixed-precision-training/index.html)\n documentation.\n* Techniques used for mixed precision training, see the [Mixed-Precision\n Training of Deep Neural\n Networks](https://devblogs.nvidia.com/mixed-precision-training-deep-neural-networks/)\n blog.\n* How to access and enable AMP for TensorFlow, see [Using\n TF-AMP](https://docs.nvidia.com/deeplearning/dgx/tensorflow-user-guide/index.html#tfamp)\n from the TensorFlow User Guide. \n\n#### Enabling mixed precision\n\nMixed precision is enabled in TensorFlow by using the Automatic Mixed Precision (TF-AMP)\nextension which casts variables to half-precision upon retrieval, while storing variables\nin single-precision format. Furthermore, to preserve small gradient magnitudes in backpropagation,\na [loss scaling](https://docs.nvidia.com/deeplearning/sdk/mixed-precision-training/index.html#lossscaling)\nstep must be included when applying gradients. In TensorFlow, loss scaling can be applied statically by\nusing simple multiplication of loss by a constant value or automatically, by TF-AMP. Automatic mixed\nprecision makes all the adjustments internally in TensorFlow, providing two benefits over manual operations.\nFirst, programmers need not modify network model code, reducing development and maintenance effort.\nSecond, using AMP maintains forward and backward compatibility with all the APIs for defining and running TensorFlow models.\n\nTo enable mixed precision, you can simply add the values to the environmental variables inside your training script:\n- Enable TF-AMP graph rewrite:\n ```\n os.environ[\"TF_ENABLE_AUTO_MIXED_PRECISION_GRAPH_REWRITE\"] = \"1\"\n ```\n \n- Enable Automated Mixed Precision:\n ```\n os.environ['TF_ENABLE_AUTO_MIXED_PRECISION'] = '1'\n ```\n\n#### Enabling TF32\n\nTensorFloat-32 (TF32) is the new math mode in [NVIDIA A100](https://www.nvidia.com/en-us/data-center/a100/) GPUs for handling the matrix math also called tensor operations. TF32 running on Tensor Cores in A100 GPUs can provide up to 10x speedups compared to single-precision floating-point math (FP32) on 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.","displayName","Transformer-XL for TensorFlow1",["D",1680632744670],"publisher","NVIDIA Deep Learning Examples","latestVersion","collectionCounts","imageUrl","/api/image/nvidia/dle/resources/transformerxl_tf1","rawImageUrl","https://assets.nvidiagrid.net/ngc/logos/Deep-Learning-Examples.png","canGuestDownload","latestVersionSizeInBytes","licenseTerms","isMultinodeEnabled","isMultiArchitecture","bias","explainability","privacy","safetyAndSecurity","featureLabels",[],"displayLabels",[273],"policyLabels",[],"systemLabels",[],"products",[],"hasNimMetadata",{"_242":274,"_275":276,"_277":278,"_279":86},"Natural Language Understanding","value","uscs_natural_language_understanding","filter",{"_280":281,"_275":276},"category","field","application",{"_150":1263,"_152":1264},"ngcApps",[1216],"csaApiBaseUrl","https://www.nvidia.com/ai-agents/csa-agent/dcsm","datadogRum",{"_1197":1198,"_1199":1200,"_1201":1202,"_1203":1204,"_1205":293,"_108":319,"_1206":1207,"_1208":59,"_1209":89,"_1210":89,"_1211":1212,"_1213":89,"_1214":1215},"amplitude",{"_1194":1195,"_831":-7,"_1196":34},"disableJs","environment","prod","featureFlagsBlob","JTdCJTIydGVzdEZsYWclMjIlM0FmYWxzZSUyQyUyMnNob3dTZWFyY2hMYWJlbFNldEZpbHRlcnMlMjIlM0FmYWxzZSUyQyUyMmVuYWJsZUNzYVdpZGdldCUyMiUzQWZhbHNlJTJDJTIyZW5hYmxlUHJvZHVjdEdyb3VwaW5nQnlTdWJzY3JpcHRpb24lMjIlM0FmYWxzZSUyQyUyMmNvbnRhaW5lclNjYW5uaW5nUmVkZXNpZ25NMiUyMiUzQWZhbHNlJTdE","featureFlagsGuarded","isDevLikeEnv","isPlaywright","isMobileDevice","categorizedOrgs",{"_1188":1189,"_1190":1191,"_1192":1193},"currentOrgContext","currentContextRoles",[],"enterpriseOrgs",[],"manageableExternalUserGroups",[],"ngcApiOrigin","ngcHelmOrigin","origin","https://catalog.ngc.nvidia.com","usesMocks","theme","year",2026,"now",1789216206580,"1.15.0","defaultSearchSuggestions",{"_1071":399,"_1072":1073,"_1074":1075,"_1076":1077},"announcements",[],"consentState",{"_356":1061,"_1062":89,"_1063":34,"_1064":34,"_1065":1066},"consentConfig",{"_338":339,"_340":341,"_342":343,"_344":345,"_346":347,"_348":349,"_350":351,"_352":353,"_354":355,"_356":357},"showContactAdmin",1789216206587,[],[],[333,334,335,336,337],{"_48":63,"_50":73,"_52":145,"_54":-5,"_56":-5,"_58":59},{"_48":63,"_50":69,"_52":145,"_54":-5,"_56":-5,"_58":59},{"_48":63,"_50":64,"_52":145,"_54":-5,"_56":-5,"_58":59},{"_48":60,"_50":61,"_52":145,"_54":-5,"_56":-5,"_58":59},{"_48":49,"_50":51,"_52":145,"_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",{"_1053":1054,"_1055":1056,"_1057":1058,"_1059":1060},"bannerAcceptLabel","Accept All","bannerRejectLabel","Reject Optional","bannerSettingsLabel","Manage Settings","preferenceSaveLabel","Save Choices","preferenceRejectLabel","Decline All","alwaysActiveText","Always Active","groups",[358,359,360,361,362],{"_363":743,"_111":744,"_210":745,"_367":89,"_368":16,"_370":746,"_372":747},{"_363":646,"_111":647,"_210":648,"_367":34,"_368":62,"_370":649,"_372":650},{"_363":554,"_111":555,"_210":556,"_367":34,"_368":57,"_370":557,"_372":558},{"_363":374,"_111":375,"_210":376,"_367":34,"_368":377,"_370":378,"_372":379},{"_363":364,"_111":365,"_210":366,"_367":34,"_368":369,"_370":371,"_372":373},"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,[496,497,498,499,500,501,502,503,504,505,506,507,508,509,510,511,512,513,514,515,516,517],[380,381,382,383,384,385,386,387,388,389,390,391,392,393,394,395,396],{"_397":491,"_242":491,"_210":399,"_370":492},{"_397":487,"_242":487,"_210":399,"_370":488},{"_397":482,"_242":482,"_210":399,"_370":483},{"_397":478,"_242":478,"_210":399,"_370":479},{"_397":473,"_242":473,"_210":399,"_370":474},{"_397":468,"_242":468,"_210":399,"_370":469},{"_397":460,"_242":460,"_210":399,"_370":461},{"_397":452,"_242":452,"_210":399,"_370":453},{"_397":447,"_242":447,"_210":399,"_370":448},{"_397":442,"_242":442,"_210":399,"_370":443},{"_397":437,"_242":437,"_210":399,"_370":438},{"_397":432,"_242":432,"_210":399,"_370":433},{"_397":427,"_242":427,"_210":399,"_370":428},{"_397":420,"_242":420,"_210":399,"_370":421},{"_397":415,"_242":415,"_210":399,"_370":416},{"_397":408,"_242":408,"_210":399,"_370":409},{"_397":398,"_242":398,"_210":399,"_370":400},"hostName","amazon-adsystem.com","",[401,402],{"_111":407,"_404":398,"_405":34,"_210":406},{"_111":403,"_404":398,"_405":34,"_210":406},"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",[410,411],{"_111":414,"_404":408,"_405":34,"_210":413},{"_111":412,"_404":408,"_405":34,"_210":413},"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",[417],{"_111":418,"_404":415,"_405":34,"_210":419},"MUID","This domain is owned by Mircosoft - it is the site for the search engine Bing.","assets.adobedtm.com",[422,423],{"_111":426,"_404":420,"_405":34,"_210":425},{"_111":424,"_404":420,"_405":34,"_210":425},"kndctr_F207D74D549850760A4C98C6_AdobeOrg_*","If you know anything about this host and its uses please add your contributions here. ","kndctr_F207D74D549850760A4C98C6_AdobeOrg_consent","tiktok.com",[429],{"_111":430,"_404":427,"_405":34,"_210":431},"_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",[434],{"_111":435,"_404":432,"_405":34,"_210":436},"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",[439],{"_111":440,"_404":437,"_405":34,"_210":441},"sp","Testing","google.com",[444],{"_111":445,"_404":442,"_405":34,"_210":446},"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",[449],{"_111":450,"_404":447,"_405":34,"_210":451},"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",[454,455],{"_111":458,"_404":452,"_405":34,"_210":459},{"_111":456,"_404":452,"_405":34,"_210":457},"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",[462,463],{"_111":466,"_404":460,"_405":34,"_210":467},{"_111":464,"_404":460,"_405":34,"_210":465},"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",[470],{"_111":471,"_404":468,"_405":34,"_210":472},"demdex","This cookie helps Adobe Audience Manger perform basic functions such as visitor identification, ID synchronization, segmentation, modeling, reporting, etc.","im-apps.net",[475],{"_111":476,"_404":473,"_405":34,"_210":477},"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",[480],{"_111":481,"_404":478,"_405":34,"_210":441},"R","6sc.co",[484],{"_111":485,"_404":482,"_405":34,"_210":486},"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",[489],{"_111":490,"_404":487,"_405":34,"_210":441},"_hly_gvid","t.contentsquare.net",[493],{"_111":494,"_404":491,"_405":89,"_210":495},"_cs_*","This is a session cookie. This is a pattern type cookie where a random 13-digit number is appended to the prefix _cs_.",{"_111":552,"_404":519,"_405":34,"_210":553},{"_111":550,"_404":519,"_405":34,"_210":551},{"_111":548,"_404":519,"_405":34,"_210":549},{"_111":430,"_404":519,"_405":34,"_210":547},{"_111":524,"_404":519,"_405":89,"_210":526},{"_111":546,"_404":519,"_405":34,"_210":399},{"_111":545,"_404":522,"_405":34,"_210":399},{"_111":543,"_404":519,"_405":34,"_210":544},{"_111":541,"_404":519,"_405":34,"_210":542},{"_111":540,"_404":519,"_405":34,"_210":399},{"_111":538,"_404":519,"_405":34,"_210":539},{"_111":524,"_404":537,"_405":89,"_210":526},{"_111":535,"_404":522,"_405":34,"_210":536},{"_111":535,"_404":525,"_405":34,"_210":536},{"_111":494,"_404":519,"_405":34,"_210":495},{"_111":533,"_404":519,"_405":34,"_210":534},{"_111":440,"_404":531,"_405":34,"_210":532},{"_111":529,"_404":519,"_405":34,"_210":530},{"_111":527,"_404":519,"_405":34,"_210":528},{"_111":524,"_404":525,"_405":89,"_210":526},{"_111":521,"_404":522,"_405":34,"_210":523},{"_111":518,"_404":519,"_405":89,"_210":520},"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.",[588,589,590,591,592,593,594,595,596,597,598,599,600,601,602,603,604,605,606,607,608,609,610,611,612],[559,560,561,562,563],{"_397":473,"_242":473,"_210":399,"_370":583},{"_397":447,"_242":447,"_210":399,"_370":578},{"_397":573,"_242":573,"_210":399,"_370":574},{"_397":568,"_242":568,"_210":399,"_370":569},{"_397":564,"_242":564,"_210":399,"_370":565},"nvidia.partners",[566],{"_111":567,"_404":564,"_405":34,"_210":441},"edge","app.hushly.com",[570],{"_111":571,"_404":568,"_405":34,"_210":572},"*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",[575],{"_111":576,"_404":573,"_405":34,"_210":577},"_octo","This domain is owned by Github. The main business activity is: Source code repository and collaboration platform",[579,580],{"_111":582,"_404":447,"_405":34,"_210":451},{"_111":581,"_404":447,"_405":34,"_210":451},"lidc","li_gc",[584,585],{"_111":587,"_404":473,"_405":34,"_210":477},{"_111":586,"_404":473,"_405":34,"_210":477},"imid_*","imid",{"_111":613,"_404":522,"_405":34,"_210":615},{"_111":644,"_404":519,"_405":34,"_210":645},{"_111":613,"_404":643,"_405":34,"_210":615},{"_111":641,"_404":519,"_405":89,"_210":642},{"_111":639,"_404":519,"_405":34,"_210":640},{"_111":638,"_404":621,"_405":89,"_210":399},{"_111":613,"_404":617,"_405":34,"_210":615},{"_111":636,"_404":519,"_405":34,"_210":637},{"_111":613,"_404":525,"_405":34,"_210":615},{"_111":624,"_404":522,"_405":34,"_210":399},{"_111":613,"_404":635,"_405":34,"_210":615},{"_111":633,"_404":519,"_405":34,"_210":634},{"_111":630,"_404":631,"_405":34,"_210":632},{"_111":628,"_404":629,"_405":34,"_210":399},{"_111":613,"_404":625,"_405":34,"_210":615},{"_111":627,"_404":519,"_405":34,"_210":399},{"_111":613,"_404":626,"_405":34,"_210":615},{"_111":624,"_404":625,"_405":34,"_210":399},{"_111":623,"_404":621,"_405":89,"_210":399},{"_111":613,"_404":622,"_405":34,"_210":615},{"_111":613,"_404":621,"_405":34,"_210":615},{"_111":613,"_404":620,"_405":34,"_210":615},{"_111":619,"_404":519,"_405":89,"_210":615},{"_111":616,"_404":617,"_405":89,"_210":618},{"_111":613,"_404":614,"_405":34,"_210":615},"_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.",[671,672,673,674,675,676,677,678,679,680,681,682,683,684,685,686,687,688,689,690,691,692,693,694,695,696,697,698,699],[651,652,653,654],{"_397":666,"_242":666,"_210":399,"_370":667},{"_397":661,"_242":661,"_210":399,"_370":662},{"_397":460,"_242":460,"_210":399,"_370":658},{"_397":573,"_242":573,"_210":399,"_370":655},[656],{"_111":657,"_404":573,"_405":89,"_210":577},"_gh_sess",[659],{"_111":660,"_404":460,"_405":89,"_210":467},"YSC","bizible.com",[663],{"_111":664,"_404":661,"_405":34,"_210":665},"_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",[668],{"_111":669,"_404":666,"_405":34,"_210":670},"tuid","This domain is owned by Ookla which provides internet speed testing services through its Speedtest application.",{"_111":741,"_404":519,"_405":34,"_210":742},{"_111":726,"_404":740,"_405":89,"_210":727},{"_111":709,"_404":522,"_405":34,"_210":634},{"_111":709,"_404":622,"_405":34,"_210":634},{"_111":738,"_404":519,"_405":34,"_210":739},{"_111":726,"_404":525,"_405":89,"_210":727},{"_111":736,"_404":519,"_405":34,"_210":737},{"_111":726,"_404":522,"_405":89,"_210":727},{"_111":734,"_404":621,"_405":34,"_210":735},{"_111":726,"_404":643,"_405":89,"_210":727},{"_111":706,"_404":522,"_405":34,"_210":708},{"_111":732,"_404":519,"_405":89,"_210":733},{"_111":730,"_404":519,"_405":34,"_210":731},{"_111":728,"_404":519,"_405":89,"_210":729},{"_111":726,"_404":617,"_405":89,"_210":727},{"_111":724,"_404":519,"_405":34,"_210":725},{"_111":722,"_404":519,"_405":34,"_210":723},{"_111":720,"_404":519,"_405":89,"_210":721},{"_111":718,"_404":626,"_405":34,"_210":719},{"_111":706,"_404":617,"_405":34,"_210":708},{"_111":716,"_404":519,"_405":34,"_210":717},{"_111":714,"_404":519,"_405":89,"_210":715},{"_111":712,"_404":519,"_405":34,"_210":713},{"_111":710,"_404":519,"_405":34,"_210":711},{"_111":709,"_404":519,"_405":34,"_210":634},{"_111":706,"_404":707,"_405":34,"_210":708},{"_111":704,"_404":519,"_405":34,"_210":705},{"_111":702,"_404":519,"_405":89,"_210":703},{"_111":700,"_404":519,"_405":34,"_210":701},"_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.",[888,889,890,891,892,893,894,895,896,897,898,899,900,901,902,903,904,905,906,907,908,909,910,911,912,913,914,915,916,917,918,919,920,921,922,923,924,925,926,927,928,929,930,931,932,933,934,935,936,937,938,939,940,941,942,943,944,945,946,947,948,949,950,951,952,953,954,955,956,957,958,959,960],[748,749,750,751,752,753,754,755,756,757,758,759,760,761,762,763,764,765,766,767,768,769,770,771,772],{"_397":885,"_242":885,"_210":399,"_370":886},{"_397":879,"_242":879,"_210":399,"_370":880},{"_397":876,"_242":876,"_210":399,"_370":877},{"_397":864,"_242":864,"_210":399,"_370":865},{"_397":861,"_242":861,"_210":399,"_370":862},{"_397":858,"_242":858,"_210":399,"_370":859},{"_397":853,"_242":853,"_210":399,"_370":854},{"_397":460,"_242":460,"_210":399,"_370":850},{"_397":847,"_242":847,"_210":399,"_370":848},{"_397":452,"_242":452,"_210":399,"_370":845},{"_397":447,"_242":447,"_210":399,"_370":843},{"_397":838,"_242":838,"_210":399,"_370":839},{"_397":833,"_242":833,"_210":399,"_370":834},{"_397":828,"_242":828,"_210":399,"_370":829},{"_397":825,"_242":825,"_210":399,"_370":826},{"_397":573,"_242":573,"_210":399,"_370":822},{"_397":432,"_242":432,"_210":399,"_370":819},{"_397":568,"_242":568,"_210":399,"_370":812},{"_397":564,"_242":564,"_210":399,"_370":806},{"_397":800,"_242":800,"_210":399,"_370":801},{"_397":795,"_242":795,"_210":399,"_370":796},{"_397":790,"_242":790,"_210":399,"_370":791},{"_397":786,"_242":786,"_210":399,"_370":787},{"_397":781,"_242":781,"_210":399,"_370":782},{"_397":773,"_242":773,"_210":399,"_370":774},"polyfill.io",[775,776],{"_111":779,"_404":773,"_405":34,"_210":780},{"_111":777,"_404":773,"_405":34,"_210":778},"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",[783],{"_111":784,"_404":781,"_405":34,"_210":785},"_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",[788],{"_111":789,"_404":786,"_405":34,"_210":441},"__cflb","instagram.com",[792],{"_111":793,"_404":790,"_405":34,"_210":794},"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",[797],{"_111":798,"_404":795,"_405":34,"_210":799},"__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",[802,803],{"_111":805,"_404":800,"_405":34,"_210":441},{"_111":804,"_404":800,"_405":34,"_210":441},"AWSALBTG","AWSALBTGCORS",[807,808],{"_111":810,"_404":564,"_405":34,"_210":811},{"_111":809,"_404":564,"_405":34,"_210":441},"ak_bmsc","bm_mi","This domain is likely associated with NVIDIA's partner program, which could involve marketing, sales, and support services for partners.",[813,814,815],{"_111":818,"_404":568,"_405":34,"_210":441},{"_111":817,"_404":568,"_405":89,"_210":441},{"_111":816,"_404":568,"_405":34,"_210":441},"AWSALBCORS","JSESSIONID","AWSALB",[820],{"_111":798,"_404":432,"_405":34,"_210":821},"This is a CloudFoundry cookie",[823],{"_111":824,"_404":573,"_405":34,"_210":577},"logged_in","hcaptcha.com",[827],{"_111":798,"_404":825,"_405":34,"_210":821},"tracking.intentsify.io",[830],{"_111":831,"_404":828,"_405":34,"_210":832},"userId","Used to track usage of the embedded video player.","static.reo.dev",[835],{"_111":836,"_404":833,"_405":34,"_210":837},"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",[840],{"_111":841,"_404":838,"_405":34,"_210":842},"dd_cookie_test_","dd_cookie_test",[844],{"_111":798,"_404":447,"_405":34,"_210":821},[846],{"_111":798,"_404":452,"_405":34,"_210":821},"supabase.co",[849],{"_111":798,"_404":847,"_405":34,"_210":821},[851],{"_111":852,"_404":460,"_405":34,"_210":467},"VISITOR_PRIVACY_METADATA","dosa-prd.azurewebsites.net",[855],{"_111":856,"_404":853,"_405":34,"_210":857},"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",[860],{"_111":798,"_404":858,"_405":34,"_210":821},"cdn.readme.io",[863],{"_111":798,"_404":861,"_405":34,"_210":821},"queue-it.net",[866,867,868,869,870],{"_111":875,"_404":864,"_405":34,"_210":399},{"_111":874,"_404":864,"_405":34,"_210":399},{"_111":873,"_404":864,"_405":34,"_210":399},{"_111":872,"_404":864,"_405":34,"_210":399},{"_111":871,"_404":864,"_405":34,"_210":399},"Queue-it-{queueId}","Queue-it-{dmonvidia }________{eventId}","Queue-it-token","queueitsoftblock_{queueId}","Queue-it-{nvidia}________{eventId}","x.com",[878],{"_111":798,"_404":876,"_405":34,"_210":799},"api.hcaptcha.com",[881,882],{"_111":884,"_404":879,"_405":34,"_210":441},{"_111":789,"_404":879,"_405":34,"_210":883},"This domain is associated with hCaptcha, a service that provides CAPTCHA and reCAPTCHA alternatives for user verification and bot protection.","hmt_id","vimeo.com",[887],{"_111":798,"_404":885,"_405":34,"_210":821},{"_111":1021,"_404":519,"_405":34,"_210":1052},{"_111":1050,"_404":614,"_405":89,"_210":1051},{"_111":1043,"_404":1012,"_405":89,"_210":1044},{"_111":1049,"_404":707,"_405":89,"_210":399},{"_111":1047,"_404":519,"_405":34,"_210":1048},{"_111":805,"_404":1033,"_405":34,"_210":972},{"_111":980,"_404":1042,"_405":89,"_210":982},{"_111":804,"_404":971,"_405":34,"_210":972},{"_111":983,"_404":1042,"_405":34,"_210":984},{"_111":1045,"_404":519,"_405":89,"_210":1046},{"_111":1043,"_404":1019,"_405":89,"_210":1044},{"_111":977,"_404":1042,"_405":34,"_210":979},{"_111":1040,"_404":519,"_405":34,"_210":1041},{"_111":804,"_404":963,"_405":34,"_210":972},{"_111":968,"_404":1039,"_405":34,"_210":970},{"_111":816,"_404":969,"_405":34,"_210":985},{"_111":818,"_404":969,"_405":34,"_210":1036},{"_111":983,"_404":978,"_405":34,"_210":984},{"_111":1037,"_404":617,"_405":34,"_210":1038},{"_111":818,"_404":643,"_405":34,"_210":1036},{"_111":1034,"_404":519,"_405":34,"_210":1035},{"_111":810,"_404":643,"_405":34,"_210":1029},{"_111":1006,"_404":622,"_405":89,"_210":399},{"_111":804,"_404":1033,"_405":34,"_210":972},{"_111":1031,"_404":971,"_405":89,"_210":1032},{"_111":995,"_404":519,"_405":34,"_210":986},{"_111":968,"_404":519,"_405":34,"_210":970},{"_111":1030,"_404":531,"_405":34,"_210":399},{"_111":810,"_404":519,"_405":34,"_210":1029},{"_111":818,"_404":963,"_405":34,"_210":1028},{"_111":1026,"_404":626,"_405":89,"_210":1027},{"_111":1024,"_404":519,"_405":89,"_210":1025},{"_111":1023,"_404":519,"_405":34,"_210":399},{"_111":1021,"_404":969,"_405":34,"_210":1022},{"_111":1020,"_404":531,"_405":34,"_210":399},{"_111":1011,"_404":1019,"_405":89,"_210":1013},{"_111":1017,"_404":622,"_405":34,"_210":1018},{"_111":1015,"_404":707,"_405":34,"_210":1016},{"_111":1014,"_404":525,"_405":34,"_210":399},{"_111":1011,"_404":1012,"_405":89,"_210":1013},{"_111":1009,"_404":978,"_405":89,"_210":1010},{"_111":1007,"_404":519,"_405":34,"_210":1008},{"_111":1006,"_404":522,"_405":89,"_210":399},{"_111":977,"_404":981,"_405":34,"_210":979},{"_111":809,"_404":996,"_405":34,"_210":986},{"_111":1004,"_404":522,"_405":34,"_210":1005},{"_111":1002,"_404":519,"_405":34,"_210":1003},{"_111":1000,"_404":519,"_405":34,"_210":1001},{"_111":999,"_404":519,"_405":34,"_210":399},{"_111":841,"_404":522,"_405":34,"_210":998},{"_111":995,"_404":643,"_405":34,"_210":986},{"_111":798,"_404":614,"_405":34,"_210":997},{"_111":995,"_404":996,"_405":34,"_210":986},{"_111":994,"_404":531,"_405":34,"_210":399},{"_111":993,"_404":519,"_405":34,"_210":966},{"_111":991,"_404":978,"_405":34,"_210":992},{"_111":804,"_404":989,"_405":34,"_210":990},{"_111":988,"_404":531,"_405":34,"_210":399},{"_111":987,"_404":519,"_405":34,"_210":399},{"_111":980,"_404":978,"_405":89,"_210":982},{"_111":809,"_404":643,"_405":34,"_210":986},{"_111":816,"_404":643,"_405":34,"_210":985},{"_111":983,"_404":981,"_405":34,"_210":984},{"_111":980,"_404":981,"_405":89,"_210":982},{"_111":977,"_404":978,"_405":34,"_210":979},{"_111":975,"_404":971,"_405":89,"_210":976},{"_111":973,"_404":629,"_405":34,"_210":974},{"_111":805,"_404":971,"_405":34,"_210":972},{"_111":968,"_404":969,"_405":34,"_210":970},{"_111":798,"_404":621,"_405":34,"_210":967},{"_111":965,"_404":519,"_405":34,"_210":966},{"_111":816,"_404":963,"_405":34,"_210":964},{"_111":961,"_404":519,"_405":34,"_210":962},"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.

",{"_743":16,"_646":59,"_554":59,"_374":59,"_364":59},"hasConsented","gpcApplied","showBanner","geolocation",{"_1067":1068,"_1069":1070},"country","US","state","OR","query","categories",[1090,1091,1092,1093,1094],"optionUrls",[1078,1079,1080,1081,1082,1083,1084,1085,1086,1087,1088,1089],"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":1168,"_1096":1169,"_1098":59},{"_20":1133,"_1096":1134,"_1098":62},{"_20":1116,"_1096":1117,"_1098":1118},{"_20":21,"_1096":1109,"_1098":1110},{"_20":1095,"_1096":1097,"_1098":19},"helm-charts","items",[1099,1100],"startIndex",{"_20":1095,"_242":1106,"_204":205,"_206":-7,"_237":1107,"_245":1103,"_270":1108,"_1105":1088},{"_20":1095,"_242":1101,"_204":205,"_206":-7,"_237":1102,"_245":1103,"_270":1104,"_1105":1089},"NVIDIA NIM Operator","k8s-nim-operator","NVIDIA",[],"url","GPU Operator","gpu-operator",[],[1111],9,{"_20":21,"_242":1112,"_204":205,"_206":1113,"_237":1114,"_245":1103,"_270":1115,"_1105":1087},"Riva Skills Quick Start","riva","riva_quickstart",[],"models",[1119,1120,1121],6,{"_20":1116,"_242":1129,"_204":205,"_206":1130,"_237":1131,"_245":1103,"_270":1132,"_1105":1084},{"_20":1116,"_242":1126,"_204":205,"_206":1123,"_237":1127,"_245":1103,"_270":1128,"_1105":1085},{"_20":1116,"_242":1122,"_204":205,"_206":1123,"_237":1124,"_245":1103,"_270":1125,"_1105":1086},"TrafficCamNet","tao","trafficcamnet",[],"PeopleNet","peoplenet",[],"StyleGAN3 pretrained models","research","stylegan3",[],"collections",[1135,1136,1137],{"_20":1133,"_242":1165,"_204":205,"_206":1139,"_237":1166,"_245":1103,"_270":1167,"_1105":1081},{"_20":1133,"_242":1162,"_204":205,"_206":-7,"_237":1163,"_245":1103,"_270":1164,"_1105":1082},{"_20":1133,"_242":1138,"_204":205,"_206":1139,"_237":1140,"_245":1103,"_270":1141,"_1105":1083},"Omniverse Kit App Streaming","omniverse","kit-appstreaming-collection",[1142,1143,1144],{"_111":1156,"_242":1157,"_239":1158,"_249":1159,"_1149":1160},{"_111":1139,"_242":1152,"_239":1153,"_249":1148,"_1149":1154},{"_111":1145,"_242":1146,"_239":1147,"_249":1148,"_1149":1150},"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",[1151],"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.",[1151,1155],"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",[1151,1155,1161],"NGC_CONSOLE","DeepStream SDK","deepstream_sdk",[],"Omniverse Kit (FB)","kit",[1142,1143,1144],"containers",[1170,1171,1172],{"_20":1168,"_242":1184,"_204":1177,"_206":1185,"_237":1186,"_245":1103,"_270":1187,"_1105":1078},{"_20":1168,"_242":1176,"_204":1177,"_206":205,"_237":1178,"_245":1103,"_270":1179,"_1105":1079},{"_20":1168,"_242":1173,"_204":205,"_206":-7,"_237":1174,"_245":1103,"_270":1175,"_1105":1080},"PyTorch","pytorch",[],"Llama-3.1-Nemotron-70B-Instruct","nim","llama-3.1-nemotron-70b-instruct",[1142,1180],{"_111":1181,"_242":1146,"_239":1147,"_249":1182,"_1149":1183},"nim-dev","https://assets.ngc.nvidia.com/products/catalog/images/product-nim-dev.jpg",[1151],"DeepSeek-R1","deepseek-ai","deepseek-r1",[1142,1180],"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",{"_1217":1218,"_1219":312,"_1220":34,"_1221":1222,"_1096":1223,"_1224":1225,"_1226":1218,"_1227":34,"_1228":34,"_1229":34,"_1230":1231},"appName","ngc-catalog","appOrigin","hideOnPrivateOrg","icon","ObjectsModules2",[1232,1233,1234,1235,1236,1237,1238],"menuIcon","editor-swatches","productName","requiresAllScopes","requiresLogin","requiresOrgLevelScopesOnly","title","Catalog",{"_1230":1260,"_50":1261,"_1224":1262},{"_1253":1254,"_1230":1255,"_50":1256,"_1226":1156},{"_1230":1251,"_50":1252,"_1224":1225},{"_1230":1248,"_50":1249,"_1224":1250},{"_1230":1245,"_50":1246,"_1224":1247},{"_1230":1242,"_50":1243,"_1224":1244},{"_1230":1239,"_50":1240,"_1224":1241},"Resources","/resources","misc-library","Models","/models","misc-neural-network","Helm Charts","/helm-charts","misc-ship-wheel","Containers","/containers","3d-cube","Collections","/collections","access",[1257],"Enterprise Catalog","/enterprise",[1258,1259],"NVAIE","READ","Explore Catalog","/","common-home",{"_155":156,"_157":158,"_159":160,"_161":162,"_163":164,"_165":166},{"_154":34}]