search

X-PLUG / mPLUG-Owl

mPLUG-Owl: The Powerful Multi-modal Large Language Model Family
https://www.modelscope.cn/studios/damo/mPLUG-Owl
MIT License
2.25k stars 171 forks source link

Can we customize the visual encoder and llm used? #69

Closed Akiraxty closed 1 year ago

MAGAer13 commented 1 year ago

You are free to replace the visual encoder and LLM. However, once you have replaced one of these, you need to pre-train it first. We recommend you to use the checkpoint we provided.

Akiraxty commented 1 year ago

Thank you. It seems that the pre-trained code has not been released yet. I haven't found it.

MAGAer13 commented 1 year ago

You can refer to the finetuning code for pre-train. The pretrain is similar to the finetuning except for we did not use any prompt.