RuntimeError: output with shape [16, 100, 100] doesn't match the broadcast shape [1, 16, 100, 100]

[Fangyi-Chen]

Hi I am running the training code and I meet the following error; I am using pytorch1.3.1 torchvision0.4.2 cuda10.1

File "AdaMixer/mmdet/models/roi_heads/adamixer_decoder.py", line 60, in _bbox_forward
featmap_strides=self.featmap_strides)
File "/home/python3.7/site-packages/torch/nn/modules/module.py", line 541, in call
result = self.forward(*input, **kwargs)
File "/home/python3.7/site-packages/mmcv/runner/fp16_utils.py", line 95, in new_func
return old_func(*args, **kwargs)
File "AdaMixer/mmdet/models/roi_heads/bbox_heads/adamixer_decoder_stage.py", line 302, in forward
attn_mask=attn_bias,
File "/home/python3.7/site-packages/torch/nn/modules/module.py", line 541, in call
result = self.forward(*input, **kwargs)
File "/home/python3.7/site-packages/mmcv/cnn/bricks/transformer.py", line 137, in forward
key_padding_mask=key_padding_mask)[0]
File "/home/python3.7/site-packages/torch/nn/modules/module.py", line 541, in call
result = self.forward(*input, **kwargs)
File "/home/python3.7/site-packages/torch/nn/modules/activation.py", line 783, in forward
attn_mask=attn_mask)
File "/home/python3.7/site-packages/torch/nn/functional.py", line 3352, in multi_head_attention_forward
attn_output_weights += attn_mask
RuntimeError: output with shape [16, 100, 100] doesn't match the broadcast shape [1, 16, 100, 100]

Looking forward to your reply and many thanks. And can you please specify your pytorch and cuda version?

[sebgao]

Please install pytorch>=1.5.0 (#12).

If you do not want to change the pytorch version, changing the line blow

AdaMixer/mmdet/models/roi_heads/bbox_heads/adamixer_decoder_stage.py

Line 294 in fd0f721

attn_bias = (iof * self.iof_tau.view(1, -1, 1, 1)).flatten(0, 1)

to

attn_bias = (iof * self.iof_tau.view(1, -1, 1, 1))

may help.