How much GPU memory is required to run the code.

AIR-THU / DAIR-RCooper

[CVPR2024] Official implementation of "RCooper: A Real-world Large-scale Dataset for Roadside Cooperative Perception"

https://openaccess.thecvf.com/content/CVPR2024/html/Hao_RCooper_A_Real-world_Large-scale_Dataset_for_Roadside_Cooperative_Perception_CVPR_2024_paper.html

55 stars 5 forks source link

How much GPU memory is required to run the code. #10

Closed CBY-9527 closed 2 months ago

CBY-9527 commented 2 months ago

Thank you very much for sharing the code. When I attempted to train the cobevt model on my 24GB 3090 GPU (using the command: python opencood_plugin/opencood/tools/train.py --hypes_yaml configs/intersection_cobevt.yaml), even with a batch size set to 1, I still encountered out-of-memory errors

leofansq commented 2 months ago

Yes, the training cost for CoBEVT is a little bit high. It requeires around 25.5G GPU memory when the batch size is set to 1.