InternVL

InternVL copied to clipboard

[Feature] MPO for other mllms

2

### Motivation 根据结构来看好像是用dpo脚本训练internVL的MPO 请问如果是QwenVL的模型是不是不支持 有没有什么方法迁移过去 ### Related resources _No response_ ### Additional context _No response_

jumbo-q

In table 13 of your paper ”Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scaling” shows that the **result of InternVL2_5-2B on GSM8K（4-shot)** is about **55**,...

lynshwoo2022

[Feature] Torchrun for MPO training

1

### Motivation Many of us only have a single node with several GPUs, and it is more common to use torchrun than srun. Hopefully, there will be an official script...

MathewCrespo

[Bug] output Lots of single "r" when inference

5

### Checklist - [x] 1. I have searched related issues but cannot get the expected help. - [x] 2. The bug has not been fixed in the latest version. -...

dongwhfdyer

Internvl2.5 for segmentation

1

Hi, is it possible to use Internvl2.5 to do segmentation tasks ?

wzczc

InternVL2_5-4B-MPO lora微调

4

在目录下没有看到MPO lora的微调，请问是目前不支持吗，还是说MPO的lora微调用的是2_5的lora微调脚本？

ChenJian7578

### Motivation Would it be possible to add a GRPO fine tuning stage to InternVL (2.5) ? I believe it would be great to teach InternVL how to reason without...

paulpacaud

[Docs] Misleading documentation of the finetuning process

1

### 📚 The doc issue In the documentation "Enhancing InternVL2 on COCO Caption Using LoRA Fine-Tuning", it is written at the end to copy the config.json file from the original...

paulpacaud

[Docs] 在mlp1层之前，新增一个全连接层，应该怎么训练？

6

### 📚 The doc issue 在mlp1层之前，新增一个全连接层，应该怎么训练，视觉层和语言层都参与训练吗？辛苦帮忙解答下@czczup ### Suggest a potential alternative/fix _No response_

DankoZhang

[Bug] FlashAttention Error During InternVL Fine-tuning on Tesla T4 GPU

3

### Checklist - [x] 1. I have searched related issues but cannot get the expected help. - [ ] 2. The bug has not been fixed in the latest version....

kachhadiyaraj15