How to use multiple GPUs for training

I2-Multimedia-Lab / A2S-v3

[IJCAI2024] The implementation of "Unified Unsupervised Salient Object Detection via Knowledge Transfer"

MIT License

14 stars 2 forks source link

How to use multiple GPUs for training #2

Open mikkker opened 2 weeks ago

mikkker commented 2 weeks ago

Hello, thank you for your excellent work. I have a question, how can I use multiple GPUs for training?

Ayews commented 2 weeks ago

Hello, thank you for your interest. Unfortunately, we did not implement multiple GPUS training in our code. You can implement the relevant parts yourself, mainly by modifying the base/framework_factory.py and train.py. However, it is worth mentioning that a single card with 24GB memory is sufficient for the training of this work, and the necessity of multiple GPUS training is not significant.