Shengqiong Wu

Hi. To my understanding, there is no fixed frame number when ImageBind encodes video. In practice, they first evenly split the video into (clips_per_video) increments and sample clips of size...

Hi @yhyu13, thx for the valuable suggestions. We have updated our code, please consult the new version.

@yhyu13, thx for the valuable feedback. - For issue 1, I have made the necessary updates to the `requirements.txt` file. - For issue 5, you are right, and you have...

Hi, we have already released the new version of the code; please refer to it.

@anonymous-atom. The weights currently on Huggingface are for the older codebase. The new checkpoint will be uploaded soon with the updated version—stay tuned!

https://github.com/NExT-GPT/NExT-GPT/blob/630b6f1b0ebd6d69d772adf31344c92514b863c4/NExT-GPT-Lagacy/code/model/anyToImageVideoAudio.py#L272 Do you mean this function?

@anonymous-atom @koookieee I just correct the error, please try again.

It's working fine on my end. What does the error log say on your side?

> Hi @ChocoWu , when the data for MoSIT will be released ? Also did you train on 2M or 10M WebVid samples ? The MoSIT can be found in...

It works fine on my side, I have no idea about the issue (⊙o⊙)…