CUDA OOM解决办法:
参考了Github上的解决思路
If you encounter a CUDA OOM error, the steps you can take to reduce your memory usage are:
1、 Reduce --batch-size
2、 Reduce --img-size
3、 Reduce model size, i.e. from YOLOv5x -> YOLOv5l -> YOLOv5m -> YOLOv5s
4、 Train with multi-GPU at the same --batch-size
5、 Upgrade your hardware to a larger GPU
6、 Train on free GPU backends with up to 16GB of CUDA memory:
CUDA Out Of Memory solutions
最新推荐文章于 2023-11-14 16:57:18 发布