Hi, thanks for your great work and for releasing the code.
I would like to continue fine-tuning LLaVA-OneVision-2, but I am a bit confused about the training code structure in this repository.
It seems that there are two different training implementations:
- aiak_training_llm
- transformers_impl
Could you please clarify which one should be used for fine-tuning LLaVA-OneVision-2 (image/video VQA tasks)?
Thanks in advance!
@anxiangsir
Hi, thanks for your great work and for releasing the code.
I would like to continue fine-tuning LLaVA-OneVision-2, but I am a bit confused about the training code structure in this repository.
It seems that there are two different training implementations:
Could you please clarify which one should be used for fine-tuning LLaVA-OneVision-2 (image/video VQA tasks)?
Thanks in advance!
@anxiangsir