I am using Faster - RCNN training network running half will appear
0000: Unable to determine the device handle for GPU 03:00. 0: the GPU is lost. Reboot the system to recover this GPU
Error, then the network card not training, before restart after retraining is no problem, but now restart after retraining will appear this problem, some people know how to solve?

CodePudding user response:

I also met

CodePudding user response:

reference 4 floor wolf_ray response:

I also met

basic can solve later I each time I restart, specific also don't know what the reason,

CodePudding user response:

As if the GPU load is too high, I four card is wide open, just hung up the only 3 card is all right,