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

关于训练DA-CLIP #91

Open
c-pupil opened this issue Dec 16, 2024 · 1 comment
Open

关于训练DA-CLIP #91

c-pupil opened this issue Dec 16, 2024 · 1 comment

Comments

@c-pupil
Copy link

c-pupil commented Dec 16, 2024

Dear author:
您好,首先感谢您出色的工作,其次我想问您一下我想按照您的方式训DA-CLIP,在我自己的训练集重头训练,我的训练集800张,用单张3090,按照下面的超参设置,每一轮大约几秒钟,显存大约15G,我想问这合理么(训练速度和占内存很令我感到意外)
python -m training.main \ --save-frequency 50 \ --zeroshot-frequency 1 \ --report-to tensorboard \ --train-data="daclip_train.csv" \ --val-data="daclip_val.csv" \ --csv-img-key filepath \ --csv-caption-key title \ --warmup 100 \ --batch-size=400 \ --lr=2e-5 \ --wd=0.05 \ --epochs=500 \ --workers=8 \ --model daclip_ViT-B-32 \ --name "daclip_ViT-B-32_b784x1_lr2e-5_e50" \ --pretrained "laion2b_s34b_b79k" \ --da

@Algolzw
Copy link
Owner

Algolzw commented Dec 18, 2024

应该是合理的,因为训练集太小而batch size太大。你可以做一些数据增强来避免过拟合。

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