💡 [REQUEST] - <支持finetune>

[dengfenglai321]

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

是否有计划支持微调？

基本示例 | Basic Example

1

缺陷 | Drawbacks

1

未解决问题 | Unresolved questions

No response

[simonJJJ]

Hi @dengfenglai321, we have supported the finetune scripts, please follow this link.

[flysnoworg]

Hi @dengfenglai321, we have supported the finetune scripts, please follow this link.

请问这种微调方式，可以让Qwen-VL-Chat学会新的图片的解读、位置定位等能力吗？因为有些业务需求的图片，Qwen-VL并不认识。
@simonJJJ

[dengfenglai321]

Hi @dengfenglai321, we have supported the finetune scripts, please follow this link.

谢谢，是否可以给出一些微调的数据集格式例子，比如图像 - 文本是怎么组织的

[Ashwath-Shetty]

is anyone able to finetune? if yes, how? any tutorials or documentations? @rujews @dengfenglai321 @simonJJJ
link mentioned is not working.

[huohuotm]

https://github.com/modelscope/swift/blob/main/examples/pytorch/llm/scripts/qwen_vl/lora_ddp/sft.sh

[ShuaiBai623]

Now we provide the official training script (full parameters, LoRA, QLoRA), finetune.py, for users to finetune the pretrained model for downstream applications in a simple fashion. Please refer to Finetuning