武帅

icon indicating a website link https://github.com/cainiaoup

icon company The University of Hong Kong icon location Hong Kong icon location A PhD student at HKU with research interests in LLM and Reinforcement Learning.

Results 5 comments of 武帅

how do i transform pth file to caffemodel?

model = yourmodel() model.load('yourmodel.pth') ## convert:

Missing mm_projector in latest LLaVa

> can you guys help in uploading pre-trained adapter file mm_projector.bin for llama-3.1-8B model ? or if available can one of you share the path pls? same question~ the projector...

Missing mm_projector in latest LLaVa

> > can you guys help in uploading pre-trained adapter file mm_projector.bin for llama-3.1-8B model ? or if available can one of you share the path pls? > > same...

How can I trian llava-next-llama-8B？

maybe you can use the finetune_clip.sh in the scripts file.

[Usage] Tokenization mismatch while finetuning Llama 3.1

you can rewrite the preprocess_llama_2 funciton follow llava-next (preprocess_llama_3)