Skip to content

Latest commit

 

History

History
44 lines (40 loc) · 1.06 KB

README.md

File metadata and controls

44 lines (40 loc) · 1.06 KB
Raw

Convert pytorch to Caffe by ONNX

This tool converts pytorch model to Caffe model by ONNX
only use for inference

Dependencies

  • caffe (with python support)
  • pytorch 0.4 (optional if you only want to convert onnx)
  • onnx

we recomand using protobuf 2.6.1 and install onnx from source

git clone --recursive https://github.com/onnx/onnx.git
cd onnx 
python setup.py install

How to use

run test.py to make sure it has been installed correctly
To convert onnx model to caffe:

python convertCaffe.py ./model/MobileNetV2.onnx ./model/MobileNetV2.prototxt ./model/MobileNetV2.caffemodel

Current support operation

  • Conv
  • ConvTranspose
  • BatchNormalization
  • MaxPool
  • AveragePool
  • Relu
  • Sigmoid
  • Dropout
  • Gemm (InnerProduct only)
  • Add
  • Mul
  • Reshape
  • Upsample
  • Concat
  • Flatten
  • Slice

TODO List

  • support all onnx operations (which is impossible)
  • merge batchnormization to convolution
  • merge scale to convolution