TensorFlow Release 21.05
The NVIDIA container image of TensorFlow, release 21.05, is available on NGC.
Contents of the TensorFlow container
This container image includes the complete source of the NVIDIA version of TensorFlow in
/opt/tensorflow. It is pre-built and installed as a system Python module.
To achieve optimum TensorFlow performance, for image based training, the container includes a sample script that demonstrates efficient training of convolutional neural networks (CNNs). The sample script may need to be modified to fit your application. The container also includes the following:
- Ubuntu 20.04
21.05-tf2-py3contains Python 3.8
- NVIDIA CUDA 11.3.0
- cuBLAS 184.108.40.206
- NVIDIA cuDNN 220.127.116.11
- NVIDIA NCCL 2.9.8 (optimized for NVLink™ )
- Horovod 0.21.3
- rdma-core 32.1
- OpenMPI 4.1.1rc1
- OpenUCX 1.10.0
- GDRCopy 2.2
- NVIDIA HPC-X 2.8.2rc3
- Nsight Compute 2021.1.0.18
- Nsight Systems 2021.1.3.14
- TensorRT 18.104.22.168
21.05-tf1-py3includes version 1.15.0+nv21.4
21.05-tf2-py3includes version TensorBoard 2.4.1
- OpenSeq2Seq at commit 8f040a49
- Included only in
- Included only in
- DALI 1.0.0
- DLProf 1.1.0
- Included only in
- Included only in
- Tensor Core optimized examples: (Included only in
- JupyterLab 2.3.1 including Jupyter-TensorBoard
Release 21.05 is based on NVIDIA CUDA 11.3.0, which requires NVIDIA Driver release 465.19.01 or later. However, if you are running on Data Center GPUs (formerly Tesla), for example, T4, you may use NVIDIA driver release 418.40 (or later R418), 440.33 (or later R440), 450.51 (or later R450), or 460.27 (or later R460). The CUDA driver's compatibility package only supports particular drivers. For a complete list of supported drivers, see the CUDA Application Compatibility topic. For more information, see CUDA Compatibility and Upgrades and NVIDIA CUDA and Drivers Support.
Release 21.05 supports CUDA compute capability 6.0 and higher. This corresponds to GPUs in the NVIDIA Pascal, Volta, Turing, and Ampere Architecture GPU families. Specifically, for a list of GPUs that this compute capability corresponds to, see CUDA GPUs. For additional support details, see Deep Learning Frameworks Support Matrix.
Key Features and Enhancements
This TensorFlow release includes the following key features and enhancements.
- TensorFlow container images version 21.05 are based on Tensorflow 1.15.5 and 2.4.0
- The environment variable TF_CUDNN_ENGINE_MAX_LIMITS can limit the number of CUDNN algos that are attempted for each convolutional layer during autotuning. This can reduce model startup costs potentially at the cost of some training throughput.
- A deterministic implementation of sparse tensor dense matmul is now available.
- Ubuntu 20.04 with April 2021 updates
- Python 2.7 is no longer supported in this TensorFlow container release.
- The TF_ENABLE_AUTO_MIXED_PRECISION environment variables are no longer supported in the tf2 container because it is not possible to automatically enable loss scaling in many cases in the tf 2.x API. Instead tf.train.experimental.enable_mixed_precision_graph_rewrite() should be used to enable AMP.
- Deep learning framework containers 19.11 and later include experimental support for Singularity v3.0.
NVIDIA TensorFlow Container Versions
The following table shows what versions of Ubuntu, CUDA, TensorFlow, and TensorRT are supported in each of the NVIDIA containers for TensorFlow. For older container versions, refer to the Frameworks Support Matrix.
|Container Version||Ubuntu||CUDA Toolkit||TensorFlow||TensorRT|
|21.05||20.04||NVIDIA CUDA 11.3.0||TensorRT 22.214.171.124|
|21.03||NVIDIA CUDA 11.2.1||TensorRT 126.96.36.199|
|21.02||NVIDIA CUDA 11.2.0||TensorRT 188.8.131.52+cuda11.1.0.024|
|20.12||NVIDIA CUDA 11.1.1||TensorRT 7.2.2|
|NVIDIA CUDA 11.1.0||TensorRT 7.2.1|
|20.09||NVIDIA CUDA 11.0.3||TensorRT 7.1.3|
|20.07||NVIDIA CUDA 11.0.194|
|20.06||NVIDIA CUDA 11.0.167|
|20.03||NVIDIA CUDA 10.2.89||TensorRT 7.0.0|
|19.10||NVIDIA CUDA 10.1.243||1.14.0|
Tensor Core Examples
The tensor core examples provided in GitHub focus on achieving the best performance and convergence by using the latest deep learning example networks and model scripts for training. Each example model trains with mixed precision Tensor Cores on Volta, therefore you can get results much faster than training without tensor cores. This model is tested against each NGC monthly container release to ensure consistent accuracy and performance over time. This container includes the following tensor core examples.
- U-Net Medical model. The U-Net model is a convolutional neural network for 2D image segmentation. This repository contains a U-Net implementation as described in the paper U-Net: Convolutional Networks for Biomedical Image Segmentation, without any alteration. This model script is available on GitHub as well as NVIDIA GPU Cloud (NGC).
- SSD320 v1.2 model. The SSD320 v1.2 model is based on the SSD: Single Shot MultiBox Detector paper, which describes an SSD as “a method for detecting objects in images using a single deep neural network”. Our implementation is based on the existing model from the TensorFlow models repository. This model script is available on GitHub as well as NVIDIA GPU Cloud (NGC).
- Neural Collaborative Filtering (NCF) model. The NCF model is a neural network that provides collaborative filtering based on implicit feedback, specifically, it provides product recommendations based on user and item interactions. The training data for this model should contain a sequence of user ID, item ID pairs indicating that the specified user has interacted with, for example, was given a rating to or clicked on, the specified item. This model script is available on GitHub as well as NVIDIA GPU Cloud (NGC).
- BERT model. BERT, or Bidirectional Encoder Representations from Transformers, is a new method of pre-training language representations which obtains state-of-the-art results on a wide array of Natural Language Processing (NLP) tasks. This model is based on BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding paper. NVIDIA's BERT is an optimized version of Google's official implementation, leveraging mixed precision arithmetic and Tensor Cores on V100 GPUS for faster training times while maintaining target accuracy. This model script is available on GitHub as well as NVIDIA GPU Cloud (NGC).
- U-Net Industrial Defect Segmentation model. This U-Net model is adapted from the original version of the U-Net model which is a convolutional auto-encoder for 2D image segmentation. U-Net was first introduced by Olaf Ronneberger, Philip Fischer, and Thomas Brox in the paper: U-Net: Convolutional Networks for Biomedical Image Segmentation. This work proposes a modified version of U-Net, called TinyUNet which performs efficiently and with very high accuracy on the industrial anomaly dataset DAGM2007. This model script is available on GitHub as well as NVIDIA GPU Cloud (NGC).
- GNMT v2 model. The GNMT v2 model is similar to the one discussed in the Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation paper. The most important difference between the two models is in the attention mechanism. In our model, the output from the first LSTM layer of the decoder goes into the attention module, then the re-weighted context is concatenated with inputs to all subsequent LSTM layers in the decoder at the current timestep. This model script is available on GitHub as well as NVIDIA GPU Cloud (NGC).
- ResNet-50 v1.5 model. The ResNet-50 v1.5 model is a modified version of the original ResNet-50 v1 model. The difference between v1 and v1.5 is in the bottleneck blocks which requires downsampling, for example, v1 has stride = 2 in the first 1x1 convolution, whereas v1.5 has stride = 2 in the 3x3 convolution. The following features were implemented in this model; data-parallel multi-GPU training with Horovod, Tensor Cores (mixed precision) training, and static loss scaling for Tensor Cores (mixed precision) training. This model script is available on GitHub as well as NVIDIA GPU Cloud (NGC).
If you encounter functional or performance issues when XLA is enabled, please refer to the XLA Best Practices document. It offers pointers on how to diagnose symptoms and possibly address them.
- A known regression can reduce the training performance of VGG-16 by up to 12% at certain batch sizes.
- Using XLA together with Horovod to parallelize training on a single node can result in out-of-memory errors. A workaround is to execute the job as follows. This will be fixed in a future release.
XLA_FLAGS=--xla_multiheap_size_constraint_per_heap=2000000000 TF_NUM_INTEROP_THREADS=1 horovodrun -np 8 bash -c 'CUDA_VISIBLE_DEVICES=$OMPI_COMM_WORLD_LOCAL_RANK python ...'
- There is a known performance regression of up to 30% when training SSD models with fp32 data type on T4 GPUs. This will be addressed in a future release.
- There is a known issue where attempting to convert some models using TF-TRT produces an error "Failed to import metagraph". This issue is still under investigation and will be resolved in a future release.
- The DLProf TensorBoard plugin included with the 21.04 and 21.05 releases is an incorrect version with respect to the DLProf command line tool included in those releases. To correct this, use the following command:
$ pip install --index-urlhttps://developer.download.nvidia.com/compute/redist nvidia_tensorboard_plugin_dlprof==1.1.0