This is the corresponding model for the paper LVChat: Facilitating Long Video Comprehension and the code (https://github.com/wangyu-ustc/LVChat).
Please download the file 7b_stage4.pth with the instructions from the github, putting the model weight under the folder ./video_models/.