choyakawa

### 起始日期 | Start Date _No response_ ### 实现PR | Implementation PR _No response_ ### 相关Issues | Reference Issues _No response_ ### 摘要 | Summary code of "needle in a...

Support LLaVA-UHD

1

https://github.com/thunlp/LLaVA-UHD This method is seemingly on par with or better than LLaVA 1.6 Next, however they opensourced the training code for reproduction. > LLM analysis from Gemini 1.5 pro: >...

support InternVL1.5

Support finetuning LLaVA 1.6

8

Support finetuning LLaVA 1.6

**Describe the feature** Please describe the feature requested here(请在这里描述需求) RLAIF-V effectively reduce the hallucination of different MLLMs **Paste any useful information** Paste any useful information, including papers, github links, etc.(请在这里描述其他有用的信息，比如相关的论文地址，github链接等)...

多模态提前将图片处理，然后再训练LLM

1

**Describe the feature** 多模态提前将图片处理，然后再训练LLM

考虑移植到新的hf transformers格式？

1

### Feature request / 功能建议 llm部分已经可以直接hf transformers了，而vision部分，似乎和glm-edge-v差别没有太多，无非是siglip和evaclip的区别 ### Motivation / 动机 方便第三方推理实现上更统一 ### Your contribution / 您的贡献 /