wickedvalley

[Question] Lora微调训练的时候报错

3

### Required prerequisites - [X] I have read the documentation . - [X] I have searched the [Issue Tracker](https://github.com/baichuan-inc/baichuan-7B/issues) and [Discussions](https://github.com/baichuan-inc/baichuan-7B/discussions) that this hasn't already been reported. (+1 or comment...

执行开启ssl提取的时候，np.isnan(ssl.detach().numpy()).sum()不为0导致后续无法训练了

7

切分的音频都是5s内，但是在执行开启ssl的时候，对应的4-cnhubert下无法生成目标文件。原因是代码执行到了： if np.isnan(ssl.detach().numpy()).sum()!= 0:return，跳出了。请问什么样的数据会导致这样的跳出，我该准备满足什么样的音频才可执行成功。

Lora微调chatglm-6b后chekpoint里面缺乏adapter_config.json文件

1

Lora微调chatglm后chekpoint里面缺乏adapter_config.json文件，这个是什么原因

[Step 0] Number of frames available for inference: 167 [Step 1] Landmarks Extraction in Video. Traceback (most recent call last): File "/docs/workspace/virtual_human/openheygen/video-retalking/inference.py", line 345, in main() File "/docs/workspace/virtual_human/openheygen/video-retalking/inference.py", line 81,...