Yinan He
Yinan He
Thank you for your careful analysis and excellent suggestions! We added support for whisper in [long_video_support](https://github.com/OpenGVLab/Ask-Anything/tree/long_video_support), we will update our code with your suggestions!
Hi，Do you sign-in with your api
https://pan.quark.cn/s/576bf2bea36f
It's already been shared with you. We will handle the mail for the first time on the Chinese workday.
Yes, we used pixel difference.
> 原来如此，不过好像在CLEVR的metadata里没有看到image_index，代码是： > > ``` > ds = datasets.load_dataset("./datasets/M3IT/", "clevr", split="train", streaming=True) > ds.info > ``` 抱歉，看到这个问题，我们是通过直接下载huggingface dataset repo里的jsonl文件读取的 
You can do so by adding your meta file in the [instruction_data](https://github.com/OpenGVLab/Ask-Anything/blob/main/video_chat2/configs/instruction_data.py) file. The meta file is a json and follows the following schema: ``` [ { "video": "string", "QA":...
you can simplely remove `bn_var_mode=syncbnVarMode_t.L2` in line94
Thank you for your feedback. You can directly comment on them, and we will fix this issue as soon as possible.