i try fuse model before QAT, and some errors occur.
"""
Swift Parameter-free Attention Network for Efficient Super-Resolution
"""
`
class SPAN(nn.Module):
In step sr = model(inputs), x size is (1, 1, 16, 16). After converting and export to onnx, x size is Tensor: (tensor(1), tensor(1), tensor(16), tensor(16)
π Describe the bug
i try fuse model before QAT, and some errors occur. """ Swift Parameter-free Attention Network for Efficient Super-Resolution """ ` class SPAN(nn.Module):
In step
sr = model(inputs)
, x size is (1, 1, 16, 16). After converting and export to onnx, x size is Tensor: (tensor(1), tensor(1), tensor(16), tensor(16)Versions
python version:
Python 3.9.13
pip install torch==2.1.2 torchvision==0.16.2 torchaudio==2.1.2 --index-url https://download.pytorch.org/whl/cu118
cc @jerryzh168 @jianyuh @raghuramank100 @jamesr66a @vkuzo @jgong5 @Xia-Weiwen @leslie-fang-intel