RuntimeError: The expanded size of the tensor (1) must match the existing size (128)

[caidwang]

I pretend to run sh ./train_minitrs.sh --gpu 0 and get several problems. All I have changed is to change device_id into device since I got a TypeError saying that TypeError: cuda() got an unexpected keyword argument 'device_id' before I get the RuntimeError lined in title.

My environment is: Unbuntu 18.04 gcc 6.4 g++6.4 cuda 9.0.176 cudnn 7.1.2 pytorch 0.4.1

[VictorLeeLk]

Hi,I got the same question,have you solved it?

[caidwang]

@VictorLeeLk Not yet, my peer changed pytorch to 0.2.0 with cuda8.0, and it works.