[tensorflow][keras]Failed to get convolution algorithm.解决方法

tenorflow-gpu 1.14或者1.13等版本经常有如下错误，这个不是你cuda安装有问题，而是显存不足造成，提示错误如下

Failed to get convolution algorithm. This is probably because cuDNN failed to initialize

 

对于tensorflow而言，代码前面加入下面代码可以解决问题

from tensorflow.compat.v1 import ConfigProto
from tensorflow.compat.v1 import InteractiveSession

config = ConfigProto()
config.gpu_options.allow_growth = True
session = InteractiveSession(config=config)

 

如果使用keras则可以用下面代码解决问题

import tensorflow as tf
import numpy as np
import keras
config = tf.compat.v1.ConfigProto(allow_soft_placement=True)
 
config.gpu_options.per_process_gpu_memory_fraction = 0.3
tf.compat.v1.keras.backend.set_session(tf.compat.v1.Session(config=config))

如果是tensorflow2.x需要用下面代码解决问题

 

import tensorflow as tf

gpus = tf.config.experimental.list_physical_devices(device_type='GPU')
for gpu in gpus:
    tf.config.experimental.set_memory_growth(gpu, True)