Spaces:
Running
on
Zero
Running
on
Zero
RuntimeError: Error(s) in loading state_dict for CLIPVisionModelWithProjection: Unexpected key(s) in state_dict: "vision_model.embeddings.position_ids".
#2
by
tatsuMura
- opened
Hi, Nice work!
I ran this code on google colab, but I met the following error in case of loading clip image encoder.
RuntimeError: Error(s) in loading state_dict for CLIPVisionModelWithProjection:
Unexpected key(s) in state_dict: "vision_model.embeddings.position_ids".
There is something wrong with image encoder weights??
Please tell me how to solve this problem.
Please check whether the package versions are right.
Thanks!
I reinstalled the package corresponding to requirements.txt and worked well.
I could resolve!
tatsuMura
changed discussion status to
closed