Replies: 1 comment
-
Also curious about this. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hello, thanks for your great work!
I ran this repo successfully and i was confused by the hack code for combine CLIPVisionModel with FSDP in https://github.com/haotian-liu/LLaVA/blob/main/llava/model/llava.py#L48
I don't think this is necessary code for FSDP. And i tried to assign model directly to vision_tower instead of loading it with list.
After that, i commented out all the code that adjusted the vision_tower device or dtype and it works.
The codes include
Is that right, or does it cause some other problem?
Beta Was this translation helpful? Give feedback.
All reactions