yw2278

Hi, I tried this rgb feature extractor, and found the result was quite different from mediapipe's

1

I try to use python to extract video features, and found this extractor, though deprecated, the model and pca params are the same as mediapipe's as I understand. However, the...

您好，感谢您的github非常有用，在使用的时候有一些疑问想跟您请教下： 1.adapter_model.bin应该是lora文件，那checkpoint-xxxx的各个文件是不是没用了，global_step里的模型很大，有什么用么 2.export_model是不是把lora和原模型融合的代码？ 3.web_demo里为什么还需要checkpoint的位置，直接是融合后的模型可以么？

[Question] 请教下原始训练数据集是否开源？

6

### 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...

lora训练完后，多轮对话会吐出其他对话

1

如图所示  训练数据是alpaca_gpt4_zh 训练方式是SFT epoch是3 想问下这种怎么解决？

有时候会出现吞字的情况

3

**Summary[问题简述（一句话）]** A clear and concise description of what the issue is. TTS生成语音的时候，会出现吞字的情况，一段文字会消失 **Env & To Reproduce[复现与环境]** 描述你用的环境、代码版本、模型 pytorch = 2.0 cpu版本 代码版本 = v0.0.1 模型 = @miven ceshi.pt **Screenshots[截图（如有）]** If...

Tooltip show event works in vega 5.25.0 but not 5.26.x

1

Please: - [x] **Describe how to reproduce the bug.** Use vega embed to plot the chart and change to devide mode in Chrome. The tooltip cannot show correctly in [email protected],...

I'm trying to test the model on local videos, and seems there is no direct way to infer on the videos.