Question: about how to fine-tune ulip to realize modelnet40 classification (not zero-shot classification)

Question

Question: about how to fine-tune ulip to realize modelnet40 classification (not zero-shot classification)

Closed this issue 16 days ago · 2 comments

Dear Dr,
I am very interested in your work, but I have a question about fine-tuning the 3d backbone on modelnet40 classification. can you give me more explanation? thank you very much!!!!!!!!!

Answer 1 · 2024-08-27T16:54:15.000Z

i guess just a simple full finetune

Answer 2 · 2024-09-15T05:34:02.000Z

Yeah, just initialize the model with ULIP pre-trained weights and possibly add the classification head. empirically I found that you need to lower the lr to get the optimal performance compared training from scratch, which makes sense since the model has been pre-trained and gained some knowledge.