TF2.0及2.1运行程序出错：Failed to get convolution algorithm. This is probably because cuDNN failed to init.._basecollectiveexecutor::startabort unknown: failed-CSDN博客

本文链接：https://blog.csdn.net/caicaiyunyun/article/details/105370179

TensorFlow2.1.0-GPU出错：Failed to get convolution algorithm

使用TF2.0或2.1。运行程序出Failed to get convolution algorithm错

运行程序出现以下错误

2020-04-06 22:36:24.284673: W tensorflow/core/common_runtime/base_collective_executor.cc:217] BaseCollectiveExecutor::StartAbort Unknown: 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 small_resnet/conv2d/Conv2D}}]]

解决问题

可能是GPU内存不足造成的。需要在程序前加以下一段代码：

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

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

不能正确导入（import）ConfigProto和InteractiveSession的解决过程

本人直接采用：

import tensorflow as tf
config = tf.compat.v1.ConfigProto()
config.gpu_options.allow_growth = True
session = tf.compat.v1.InteractiveSession(config=config)