python - The model doen't fit withTensorflow and Keras Error with GPU

Question

I installed cuda 11.0 and cudnn 8.0 with tensorflow-gpu 2.4.0. When I fit the model, I get this error.

UnknownError:  Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above.
     [[node sequential/conv2d/Relu (defined at <ipython-input-23-397fbbe99754>:1) ]] [Op:__inference_train_function_1671]

Function call stack:
train_function

I tried to install keras-gpu but the all environment and tensorflow doesn't work.

question from:https://stackoverflow.com/questions/66046177/the-model-doent-fit-withtensorflow-and-keras-error-with-gpu

Answer 1

import tensorflow as tf
gpu = tf.config.experimental.list_physical_devices('GPU')
print("Num GPUs Available: ", len(gpu))
tf.config.experimental.set_memory_growth(gpu[0], True)

This solves the error.