Why use anaconda, in accordance with the tensorflow gpu (version for 1.13.1 CUDA9)
Trained model, the occupancy rate has been 100% CPU, while GPU occupancy rate is 0%, basic
Perform a
Sess=tf.. The Session (config=tf ConfigProto (log_device_placement=True))
Code, the following result, could you please tell me the situation is in use gpu to run?

CodePudding user response:

Is to use GPU, GTX 1050 Ti

CodePudding user response:

No error, and has given a 1.3 G of GPU memory
Seems run normal

Can open in the code to run a command window to run GPU nvidia - smi to view the status

CodePudding user response:

Do you solved? I also am this problem, run the program or use of the CPU

CodePudding user response:

Model code set to use the GPU option turned on, not just add on

CodePudding user response:

Solved? Excuse me, I am also this kind of situation

CodePudding user response:

Suggested above tensorflow1.15 version, there is no need for the GPU and the GPU version, run the program with the following statement before testing can find GPU
The import tensorflow as tf
Print (" Num GPUs Available: ", len (tf) config). Experimental list_physical_devices (' GPU)))