GPU utilization rate is low

sthalles / SimCLR

PyTorch implementation of SimCLR: A Simple Framework for Contrastive Learning of Visual Representations

https://sthalles.github.io/simple-self-supervised-learning/

MIT License

2.21k stars 458 forks source link

GPU utilization rate is low #15

Closed LiJunnan1992 closed 3 years ago

LiJunnan1992 commented 4 years ago

Hi, thanks for the code!

When I tried to run it on single GPU (v-100), the utilizaiton rate is very low (~0-10%) even if I increase num_worker. Would you know why this happens and how to solve it? Thanks!

JosephElHachem commented 3 years ago

I am having the same observation. I am tempted to say that the most time consuming operations are generating the data augmentations. I haven't done the measurements yet though.

sthalles commented 3 years ago

Hello guys, that is a general message to say that I have refactored the whole project. I believe the project is much easier to understand now. Please have a look at the new impl and free to submit PR if you find any bugs. Thanks.