py3.6 pytorch tensorrt v7 onnx onnxsim pycuda
pip install pycuda
```shell
pip install onnx
pip install onnx-simplifier
```
Notably, the trt version should be compatible with the py version:
- Downloading tensorrt in 'https://developer.nvidia.com/nvidia-tensorrt-7x-download'
- tar xzvf xxx & cd TensorRT-7.x.x.x/python & pip install tensorrt-7.x.x.x-cpxx-none-linux_x86_64.whl
-
convert pth to onnx: changing config file in config.py modifying the function of load_model pth2onnx/pth_to_onnx.py
python pth2onnx/pth_to_onnx.py python pth2onnx/inference_onnx.py
-
testing the onnx
python pth2onnx/inference_onnx.py
-
convert onnx to trt: changing config file in config.py modifying the function of load_model onnx2trt/onnx2trt.py
python onnx2trt/onnx_to_trt.py python onnx2trt/inference_trt.py
-
testing the trt
python onnx2trt/inference_trt.py