fully customizable and compilable tensorflow development environment in docker
- protobuf == 3.19.1
- onnx == 1.10.2
- onednn == 2.4.3
- opencv == 4.5.4
- tensorrt == 8.0.3.4
- bazel == 3.7.2
- tensorflow == 2.7.0
- onnxruntime (wheel install)
- tensorflow-addons == 0.15.0
- tensorflow-io == 0.22.0
- and more ...
- tensorrt 8.0.3.4 from NVIDIA website: nv-tensorrt-repo-ubuntu2004-cuda11.3-trt8.0.3.4-ga-20210831_1-1_amd64.deb
- Compiled with AVX2/AVX512 on
- Compiled with highest NVIDIA compute capability supported by both device and tensorflow
- training 75% faster than vanilla wheel install (HDRNet with Intel Xeon Silver and 2080Ti)