vision-cair / chatcaptioner Goto Github PK

Thank you for your great work!

I am currently having trouble receiving any response from ChatGPT during the caption.ipynb inference process, although I am able to use ChatGPT through a web browser. Could you please let me know the expected time for each chat process during inference, and if using ChatGPT Plus would help to accelerate this process.

Thank you for your assistance!

Thank you for the great work!
When I run the code, something like the following picture appears. Is something wrong?

Hi, I can run the demo url and be able to upload an image with the first answer from BLIP. However, then the process stuck on waiting for the chatgpt reply, any ideas about this issue?

或者说有没有如何使用个人上传视频完成字幕生成的详细操作指引呢？如果能像图像字幕那样开发gradio web页面最好不过了，再次感谢开发者的工作

Thank you for the great work!
I understand to run BLIP2, I need a 24G GPU, but now I only have 4 16G GPUs, is there a way I can run this repo on multiple GPUs?

Hi,Thank you very much for your contribution！When I run the code, blip2 doesn't answer the question. As a result, the rest of the dialogue is completely blank.

Hi thanks for the work. I would like to know why is the number of frames hard coded in video_reader.py and captioning files? How can it be made dynamic so it accepts any input like a video of 3 minutes? Does the code base supports it?