TensorFlow Release 21.04

TensorFlow Release 21.04 (PDF)

The NVIDIA container image of TensorFlow, release 21.04, 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:

Driver Requirements

Release 21.04 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.

GPU Requirements

Release 21.04 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 21.04 are based on Tensorflow 1.15.5 and 2.4.0
  • Ubuntu 20.04 with March 2021 updates
  • Improved performance by caching the compilation result after LLVM IR creation and removing subsequent LLVM and PTXAS compilation phases.
  • Added GPU-deterministic tf.sparse.sparse_dense_matmul support (for the tf.float32 data type). When TF_DETERMINISTIC_OPS is set to "true" or "1" then tf.sparse.sparse_dense_matmul will operate deterministically in both the forward and backward direction.
  • Integrated CUDNN v8 API for RNN and fused conv+bias+activation ops.
  • Fixed an issue that caused OOM errors in some cases when using a batch size of 1.
  • Improved XLA handling of dynamic ops to avoid frequent recompilation.
  • Implemented XLA persistent cache.
  • Implemented custom learning rate support in Horovod.


  • 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.

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.

Known Issues


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.

  • 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.
  • There is a known CUDNN performance regression affecting certain batch sizes of VGG based models by up to 45%. This will be fixed in a later release.
  • The DLProf TensorBoard plugin included with the 21.04 release 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

© Copyright 2024, NVIDIA. Last updated on Apr 5, 2024.