CUDA kernel errors might be asynchronously reported at some other API call,so the stacktrace below m

报错信息：

RuntimeError: FBOCCFuse: OccHead: CustomFocalLoss: CUDA error: out of memory
CUDA kernel errors might be asynchronously reported at some other API call,so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.

问题及解决：CUDA在其他地方有被占用的情况，先关闭其他程序中的cuda占用，再执行。
运行ps命令查看自己的进程情况及gpu占用情况，kill -9 PID, 可以杀死对应的PID进程。


ps -ef | grep occ | cut -c 9-15 | xargs kill -9 #一次删除所有带python的进程