Visual Question Answering
Transformers
PyTorch
internvl_chat
feature-extraction
custom_code

Which exact version of Llama 2 was used?

#1
by floschne - opened

Hi, and thanks for this fantastic work!

Can you please tell me the exact version of Llama2 that was used in this model version? In the paper, it says Vicuna 13B, but here it says Llama 2 13B. Also, if it's Llama2, do you use the instruction-tuned version or the ''raw'' version? :)

OpenGVLab org

Hi, the LLM used in this model is not Vicuna, but a version that we created ourselves using the base model of LLaMA2-13B with supervised fine-tuning.

czczup changed discussion status to closed

Sign up or log in to comment