Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

training time #1

Open
SISTMrL opened this issue Jun 16, 2022 · 1 comment
Open

training time #1

SISTMrL opened this issue Jun 16, 2022 · 1 comment

Comments

@SISTMrL
Copy link

SISTMrL commented Jun 16, 2022

Hello, could you please tell me the training time of hico and gqa in which gpu, thanks!

@jeasinema
Copy link
Contributor

Hi,

Our GQA experiment took ~24h on 64 16G V100 GPUs with default hyperparameters. This is because the training is end-to-end (both MCAN and backbone would be trained jointly). But you may try to pretrain the backbone with the RelViT objective and make it frozen during the training of MCAN. Some performance degradation is expected tho.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants