This is modified from the Onnx2Caffe to support new operators, and serve for Vitis AI FPGA project.

Use the command to preproduce the conda env:

conda env create -f environment.yml -n cc_onnx2caffe

To test if the setup is successful, try to convert the provided sample model model/atss_lite0.onnx.

1. To convert onnx model to caffe, simply use:

ONNX_FILE=model/atss_lite0.onnx
python convertCaffe.py  $ONNX_FILE

• The results (model.prototext and model.caffemodel) will be saved to the same folder of ONNX file.
• If you want to save Caffe files to different folder, specify --ouput $OUTPUT_DIR.

2. To verify if the converted caffe yields the same output with ONNX model, use:

ONNX_FILE=model/atss_lite0.onnx
CAFFE_CKPT=model/atss_lite0.caffemodel 
python tools/verify_caffe_model.py $ONNX $CAFFE_CKPT --shape 1280 768

• The Mean Absolute Error(MAE) and Relative Error will be computed using a random input image.
• The shape input must be set correctly. If wrong, it will print the expected size. This is intentially to ensure you double check the correct size for inference.
• To test with specific image, set --input_img $IMAGE_FILE.

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