tensorflow-gpu2.0报错:Failed to get convolution algorithm. This is probably because cuDNN...

tensorflow-gpu2.0报错:Failed to get convolution algorithm. This is probably because cuDNN…

发现tensorflow2.0-gpu版在运行时可以使用GPU运算全连接层,不可以运行卷积运算Conv2D,报错:

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.
在网上试了很多方法,都不可行,例如动态给gpu分配内存啥的,最后按照这个方法成功解决了,即安装了自己CUDA适应的最新的cuDNN,将cuDNN7.4.5版本换成了7.6.5版本:

链接: https://blog.csdn.net/qq_41185569/article/details/107192347

注:可以在任务管理器的这个地方查看gpu在模型计算上的使用率:

你可能感兴趣的:(Python,深度学习,机器学习,tensorflow,深度学习,卷积神经网络)