Error loading checkpoint

Question

Error loading checkpoint

TalalWasim opened this issue 2 years ago · 3 comments

Hi,

I am having trouble when evaluating pretrained checkpoint. I get the following error:

RuntimeError: Error(s) in loading state_dict for STCATNet:
size mismatch for ground_encoder.encoder.time_embed.te: copying a param with shape torch.Size([301, 1, 256]) from checkpoint, the shape in current model is torch.Size([201, 1, 256]).

yoon28 commented a year ago

+1

Answer 1 · 2023-09-06T09:51:27.000Z

@jy0205 How can I resolve this problem?

Answer 2 · 2023-09-14T10:40:38.000Z

@TalalWasim
I finally figured it out.
Set INPUT.MAX_VIDEO_LEN to 300 when you execute the code.