Export to ONNX

[michaelnguyen11]

Hi authors and communities,

Do anyone convert successfully DeepPruner / DeepPruner Fast Pytorch to ONNX model ? Could you guys help me to convert DeepPruner Pytoch to ONNX model ?

I tried to convert to ONNX, however, the uniform operator is not supported with all version of ONNX opset.
How I did : in submission_kitti.py , at function test()

def test(imgL, imgR):
    model.eval()
    with torch.no_grad():
        imgL = Variable(torch.FloatTensor(imgL))
        imgR = Variable(torch.FloatTensor(imgR))

        if args.cuda:
            imgL, imgR = imgL.cuda(), imgR.cuda()


        onnx_file = f"deepPruner_fast.onnx"
        torch.onnx.export(model,
                        args=(imgL,imgR),
                        f=onnx_file,
                        opset_version=13)
        import onnx
        from onnxsim import simplify
        model_load = onnx.load(onnx_file)
        model_simp, check = simplify(model_load)
        onnx.save(model_simp, onnx_file)
        import sys
        sys.exit(0)

        refined_disparity = model(imgL, imgR)
        return refined_disparity

Many thanks in advance.

[ShivamDuggal4]

Hi @michaelnguyen11

Sorry, we do not have the ONNX version of the DeepPruner models and are not aware of anyone who can provide quick help for the same.

Best Regards
Shivam