facebookresearch / facebookresearch/SlowFast
about FC layer output size
Open
question
- Dominant language
- Python
- Stars
- 7.4k
- Forks
- 1.3k
- PR merge metrics
- No merged PRs in 30d
Description
in head_helper.py ,when i training ,after x = self.projection(x),i get x size is sub_batch * 2 * 2*2*1,after x.view (x.shape[0], -1),x size is sub_batch*8, is right?why is it?
sub_batch = batch_size // gpu_num
thx
Contributor guide
Assessment
This issue has not been assessed yet.