Kaldi Release 19.04
The NVIDIA container image for Kaldi, release 19.04, is available on NGC.
Contents of the Kaldi container
- Ubuntu 16.04 including Python 3.5
- NVIDIA CUDA 10.1.105 including cuBLAS 10.1.0.105
- NVIDIA cuDNN 7.5.0
- NVIDIA NCCL 2.4.6 (optimized for NVLink™ )
- OpenMPI 3.1.3
- TensorRT 5.1.2
Release 19.04 is based on CUDA 10.1, which requires NVIDIA Driver release 418.xx.x+. However, if you are running on Tesla (Tesla V100, Tesla P4, Tesla P40, or Tesla P100), you may use NVIDIA driver release 384.111+ or 410. 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.
Release 19.04 supports CUDA compute capability 6.0 and higher. This corresponds to GPUs in the Pascal, Volta, and Turing 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
- Kaldi container image version 19.04 is based on Kaldi 5.5.
- Latest version of NVIDIA NCCL 2.4.6
- Latest version of cuBLAS 10.1.0.105
- Reduced memory usage and improved performance
- Improved scripts for running benchmarks
- Ubuntu 16.04 with March 2019 updates
/workspace/nvidia-examples/librispeech/prepare.shwhich downloads a trained model and data.
/workspace/nvidia-examples/librispeech/run_multigpu_benchmark.shwhich run inference on the trained model and data.
There are no known issues in this release.