End2End ChatBOT for video and image with explicit communication with ChatGPT.
The Ask-Anything project has released a VideoChat tool that uses End2End ChatBOT for video and image communication. The tool has undergone instruction tuning and now supports longer videos and incorporates Langchain and Whisper. The project is ongoing, with the team studying general video understanding and long-term video reasoning, including AIGC for video. The team also welcomes researchers, engineers, and interns to join them at the General Vision Group, Shanghai AI Lab.