search

RenShuhuai-Andy / TimeChat

[CVPR 2024] TimeChat: A Time-sensitive Multimodal Large Language Model for Long Video Understanding
https://arxiv.org/abs/2312.02051
BSD 3-Clause "New" or "Revised" License
267 stars 23 forks source link

Demo can‘t show the same desult #15

Closed xiaoxiaoli666 closed 5 months ago

xiaoxiaoli666 commented 5 months ago

hello,thank you for your contribution. I met a question when run the demo.demo.ipynb,as the result can‘t be like yours. It's so short.how can I get a long and detailed description?also, the demo can't run on gpu,it run on cpu. The environment is as your readme.How can I put it on gpu? thank you!

RenShuhuai-Andy commented 5 months ago

Hi, thanks for your interest.

The following line in demo.ipynb will put the model to gpu0: image make sure your torch supports gpus (you can run python -c"import torch; print(torch.cuda.is_available())" to verify this, the expected output should be True).

For the output, what does the output on your device look like? Or can you just re-run the inference multiple times to see if the output is long and detailed?