TensorRT Release 18.07
The NVIDIA container image of TensorRT, release 18.07, is available.
Contents of TensorRT
- The TensorRT documentation and C++ samples. The samples can be built by running make in the /workspace/tensorrt/samples directory. The resulting executables are in the /workspace/tensorrt/bin directory.
- The TensorRT Python examples. The Python examples can be found in the /workspace/tensorrt/python/examples directory. Most Python examples can be run using python <script.py> /workspace/tensorrt/python/data. The Python API documentation can be found in the /usr/lib/python2.7/dist-packages/docs directory.
- Ubuntu 16.04
- NVIDIA CUDA 9.0.176 (see Errata section and 2.1) including CUDA® Basic Linear Algebra Subroutines library™ (cuBLAS) 9.0.425
- NVIDIA CUDA® Deep Neural Network library™ (cuDNN) 7.1.4
- NCCL 2.2.13 (optimized for NVLink™ )
Key Features and Enhancements
- TensorRT container image version 18.07 is based on TensorRT 4.0.1.
- Latest version of CUDA® Basic Linear Algebra Subroutines library™ (cuBLAS) 9.0.425.
- Ubuntu 16.04 with June 2018 updates
Known Issues
Some samples require data files that are not included within the TensorRT container either due to licensing concerns or because they are too large. Samples which do not include all the required data files instead include a README.txt file in the corresponding source directory informing you how to obtain the necessary data files. The data files required for the samples sampleNMT and sampleUffSSD cannot be easily created within the TensorRT container using the default packages. You should instead prepare the data files for these samples outside the container and then use docker cp to copy the necessary files into the TensorRT container or use a mount point when running the TensorRT container.