ShuaiBai623 comments

Results 53 comments of


                                            ShuaiBai623

train question

What error?

Question about Cross-MFE and Self-MFE

1 In our final implementation. Self-MFE takes warped source features and warped reference features as inputs as, and it achieves slightly better performance 2 Since the network regresses the residual...

The training result is blank

Are there screenshots of the training process?

Training model test results

Is the training process normal?

[BUG] RuntimeError: a view of a leaf Variable that requires grad is being used in an in-place operation.<title>

感谢您的尝试，在inplace前先clone()就可以解决这个问题, 可以同步下huggingface上最新的模型代码，可以参考这个https://huggingface.co/Qwen/Qwen-VL-Chat/commit/e71876524333aa79e65b19c557c639175552e689

请问如何让程序并行执行，用到多GPU卡

如果是多卡测试，可以参考eval_mm/evaluate_caption.py 的实现来进行多卡和组batch执行

纯文本指令数据和多模态指令数据混在一起finetune

可以同步下huggingface上最新代码，可以解决纯文本和图像的联合训练

纯文本指令数据和多模态指令数据混在一起finetune

有什么新错误吗

纯文本指令数据和多模态指令数据混在一起finetune

![image](https://github.com/QwenLM/Qwen-VL/assets/43326198/c0654027-d30b-44c1-b5ea-3161961a246a) hi @ZhihuaGao , 这几个更新是针对混合训练的

💡 [REQUEST] - About Qwen 14B support

感谢支持，会有的，还请等待