How to convert pytorch to onnx?

[jeromezyb]

tools/pytorch2onnx.py

python tools/pytorch2onnx.py configs/swin/upernet_swin_tiny_patch4_window7_512x512_160k_ade20k.py

Error: Exporting the operator roll to ONNX opset version 11 is not supported. Please open a bug to request ONNX export support for the missing operator.

[romil611]

I'm also looking to convert the model to onnx. If possible please suggest alternate for roll operator. thanks.

[wanchunwu-gorilla]

Add the following code to pytorch/torch/onnx/symbolic_opset11.py:

@parse_args('v', 'is', 'is')
def roll(g, self, shifts, dims):
    assert len(shifts) == len(dims)

    result = self
    for i in range(len(shifts)):
        shapes = []
        shape = sym_help._slice_helper(g,
                                       result,
                                       axes=[dims[i]],
                                       starts=[-shifts[i]],
                                       ends=[maxsize])
        shapes.append(shape)
        shape = sym_help._slice_helper(g,
                                       result,
                                       axes=[dims[i]],
                                       starts=[0],
                                       ends=[-shifts[i]])
        shapes.append(shape)
        result = g.op("Concat", *shapes, axis_i=dims[i])

    return result

Reference: pytorch/pytorch#68974

[Kaunain26]

Add the following code to pytorch/torch/onnx/symbolic_opset11.py:

@parse_args('v', 'is', 'is')
def roll(g, self, shifts, dims):
    assert len(shifts) == len(dims)

    result = self
    for i in range(len(shifts)):
        shapes = []
        shape = sym_help._slice_helper(g,
                                       result,
                                       axes=[dims[i]],
                                       starts=[-shifts[i]],
                                       ends=[maxsize])
        shapes.append(shape)
        shape = sym_help._slice_helper(g,
                                       result,
                                       axes=[dims[i]],
                                       starts=[0],
                                       ends=[-shifts[i]])
        shapes.append(shape)
        result = g.op("Concat", *shapes, axis_i=dims[i])

    return result

Reference: pytorch/pytorch#68974

this did not work. :(