Fail to find the dnn implementation node CudnnRNN

tensorflow.python.framework.errors_impl.UnknownError:    Fail to find the dnn implementation.
     [[{{node CudnnRNN}}]]
     [[model/lstm/PartitionedCall]] [Op:__inference_train_function_3343]

Function call stack:
train_function -> train_function -> train_function

解决方法一：

# 设置gpu内存自增长
gpus = tf.config.experimental.list_physical_devices('GPU')
print(gpus)
for gpu in gpus:
    tf.config.experimental.set_memory_growth(gpu, True)
# 设置gpu内存自增长
gpus = tf.config.experimental.list_physical_devices('GPU')
print(gpus)
for gpu in gpus:
    tf.config.experimental.set_memory_growth(gpu, True)

解决方法二：

安装好了Anaconda3—后，运行开始菜单—>Anaconda3—>Anaconda Prompt

conda install keras

## CPU
pip3 install tensorflow -i https://pypi.tuna.tsinghua.edu.cn/simple/

## GPU
pip3 install tensorflow-gpu -i https://pypi.tuna.tsinghua.edu.cn/simple/