openai / CLIP

CLIP (Contrastive Language-Image Pretraining), Predict the most relevant text snippet given an image

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

How to only fine-tune the image encoder?

GA-17a opened this issue · comments

When I want to feed only the parameters of the image encoder (self.model.encode_image.parameters()) into the optimizer, it shows that "AttributeError: 'function' object has no attribute 'parameters'". It's ok for "self.model.parameters()".

Anyone knows how to solve this problem?

man,cause that's a function not A module,try with the "self.model.visual"

man,cause that's a function not A module,try with the "self.model.visual"
It has been solved in the same way, thank you.

When I want to feed only the parameters of the image encoder (self.model.encode_image.parameters()) into the optimizer, it shows that "AttributeError: 'function' object has no attribute 'parameters'". It's ok for "self.model.parameters()".

Anyone knows how to solve this problem?

I met the same question,bro. Could u please help me about the solution you took