I want to transfer my torch file to onnx format, but some warning occur during the transferring:
WARNING: The shape inference of custom::deform_conv2d type is missing, so it may result in wrong shape inference for the exported graph. Please consider adding it in symbolic function.
my code:
import torch
import torchvision
class Model(torch.nn.Module):
def __init__(self):
super().__init__()
self.conv1 = torch.nn.Conv2d(3, 18, 3)
self.conv2 = torchvision.ops.DeformConv2d(3, 3, 3)
def forward(self, x):
return self.conv2(x, self.conv1(x))
from torch.onnx import register_custom_op_symbolic
from torch.onnx.symbolic_helper import parse_args
@parse_args("v", "v", "v", "v", "v", "i", "i", "i", "i", "i", "i", "i", "i", "none")
def symbolic(g,
input,
weight,
offset,
mask,
bias,
stride_h, stride_w,
pad_h, pad_w,
dil_h, dil_w,
n_weight_grps,
n_offset_grps,
use_mask):
return g.op("custom::deform_conv2d", input, offset)
register_custom_op_symbolic("torchvision::deform_conv2d", symbolic, 9)
model = Model()
input = torch.rand(1, 3, 10, 10)
torch.onnx.export(model, input, 'dcn.onnx')
I have try to setType, but not work. I hope my transferred process of '.onnx' file without warning.(Casue I will transfer 'onnx' to 'paddle' format, wrong shape inference bring a lot mistakes)
