please fix llava 70B config
Open
- Dominant language
- Python
- Stars
- 5.2k
- Forks
- 448
- Avg merge
- 3d 15h
- Merged PRs (30d)
- 26
Description
There is one pretraining config provided for llava based on llama 70B at `xtuner/configs/llava/llama3_70b_instruct_clip_vit_large_p14_336`. However, inside this config, there is 8B model pointed out
```
# Model
llm_name_or_path = 'meta-llama/Meta-Llama-3-8B-Instruct'
visual_encoder_name_or_path = 'openai/clip-vit-large-patch14-336'
```
could you please assist me how to setup correctly LlaVA pretraining and finetuning based on 70B llama model?
Contributor guide
Assessment
This issue has not been assessed yet.