Open rebuttalpapers opened 3 months ago
We use Qwen2-7B-Instrcut.
I just add a license section in our HF repo.
For the data I saw "We trained our model using the same data recipe and two-stage training approach as LLaVA-1.6."
May I ask what data you use during each stage?
Hi @rebuttalpapers , you can refer to the LLaVA-1.6 blog post: https://llava-vl.github.io/blog/2024-01-30-llava-next/
what is the base model for "lmms-lab/LongVA-7B-DPO" and "lmms-lab/LongVA-7B" Can it be used for commercial user case?