Support InternVL finetune

[hhaAndroid]

基本说明

• ./v1_5/ 包含着 InterVL 1.5 训练配置的配置文件
• 支持 InternVL 2B/4B/26B 模型全量/LoRA/QLoRA 微调，综合考虑效率性能，建议你优先选择 4B 模型
• 在训练完成后，可以使用 ./v1_5/convert_to_official.py 脚本将 XTuner 训练的模型转换为官方格式，从而复用官方所支持的所有工具链
• 目前所有配置都是以 8xA100 80G 显卡为基准，2B/4B 可以使用 ZERO1 训练，26B 模型必须要 ZERO3 运行，并且没有对参数进行过多的调整，你可以按照你自己的需求进行修改
• 目前是以 LLaVA SFT 数据进行验证，无法充分反应微调性能，你可以根据自己的需求进行数据自定义，后续我们会提供一个相对公平的微调数据集

Basic Introduction

• ./v1_5/ contains the configuration files for training InterVL 1.5
• Support InternVL 2B/4B/26B model full/LoRA/QLoRA fine-tuning, considering efficiency and performance, it is recommended to choose the 4B model first
• After training, you can use the ./v1_5/convert_to_official.py script to convert the model trained by XTuner to the official format, so as to reuse all the official supported toolchains
• All configurations are based on 8xA100 80G graphics cards, 2B/4B can use ZERO1 training, 26B models must run ZERO3, and there is no excessive adjustment of parameters, you can modify them according to your own needs
• It is verified with LLaVA SFT data, which cannot fully reflect the fine-tuning performance. You can customize the data according to your own needs. We will provide a relatively fair fine-tuning dataset later

[KooSung]

@hhaAndroid hello，请问这个进度大概完成了多少？什么时候会merge到主分支呢？

[hhaAndroid]

@hhaAndroid hello，请问这个进度大概完成了多少？什么时候会merge到主分支呢？

已经完成了，你可以直接用，这个就已经是最终形态了

[cydiachen]

find one error