search

mbzuai-oryx / Video-ChatGPT

[ACL 2024 🔥] Video-ChatGPT is a video conversation model capable of generating meaningful conversation about videos. It combines the capabilities of LLMs with a pretrained visual encoder adapted for spatiotemporal video representation. We also introduce a rigorous 'Quantitative Evaluation Benchmarking' for video-based conversational models.
https://mbzuai-oryx.github.io/Video-ChatGPT
Creative Commons Attribution 4.0 International
1.17k stars 102 forks source link

can I train in one A100 80G GPU? #51

Closed hkuit closed 1 year ago

hkuit commented 1 year ago

Hello, thanks for the great work.

Can I train the model using only one A100 80G GPU? Or how can we modify the code so that it can be trained on one gpu? Thank you so much.

mmaaz60 commented 1 year ago

Hi @hkuit,

Yes, you can train on a single A100-80G GPU. Please make sure to keep the overall batch size to 32. This can be achieved by using the following setting,

--per_device_train_batch_size 8 \
--gradient_accumulation_steps 4 \

Please let me know if it works. Thanks

hkuit commented 1 year ago

Thank you for the reply @mmaaz60 , let me try it and update to you later.

hkuit commented 1 year ago

Thanks, it works.