how could i do, the train process will set on GPUS?

[GnJoeln]

I have 2 3090 gpus, so I set gpus=[0, 1] in configs, but the time of train one epoch need 4 hours, and the output shows that Root Member(assigned_device_name_index_=-1 requested_device_name_='/device:GPU:1' assigned_device_name_='' resource_device_name_='/device:GPU:1' supported_device_types_=[CPU] possible_devices_=[]
Identity: CPU
Switch: CPU
AssignSubVariableOp: CPU
Mul: CPU
VarHandleOp: CPU
Const: CPU
VarIsInitializedOp: CPU
AssignVariableOp: CPU
ReadVariableOp: CPU
Sub: CPU
So, I guess I don't use gpu to train the model and I want to know how could run train.py on GPU? thank you for your help.

[zlc213]

您好请问您复现成功了吗