Keras中的CuDNNLSTM和LSTM有什么区别? [英] What is the difference between CuDNNLSTM and LSTM in Keras?

问题描述

在高级深度学习库Keras中，存在多种类型的循环层；其中包括LSTM(长期短期记忆)和CuDNNLSTM.根据 Keras文档，CuDNNLSTM是:

In Keras, the high-level deep learning library, there are multiple types of recurrent layers; these include LSTM (Long short term memory) and CuDNNLSTM. According to the Keras documentation, a CuDNNLSTM is a:

由CuDNN支持的快速LSTM实现. 只能在具有TensorFlow后端的GPU上运行.

Fast LSTM implementation backed by CuDNN. Can only be run on GPU, with the TensorFlow backend.

我相信Keras会尽可能自动使用GPU.根据 TensorFlow构建说明，要使TensorFlow GPU后端正常工作，您将需要CuDNN:

It is my belief that Keras automatically uses the GPU wherever possible. According to the TensorFlow build instructions, to have a working TensorFlow GPU backend, you will need CuDNN:

您的系统上必须安装以下NVIDIA软件:

The following NVIDIA software must be installed on your system:

• NVIDIA的Cuda工具包(> = 7.0).我们建议使用9.0版.有关详细信息，请参阅NVIDIA文档.确保按照NVIDIA文档中的说明将相关的Cuda路径名附加到LD_LIBRARY_PATH环境变量中.
• 与NVIDIA Cuda Toolkit相关的NVIDIA驱动程序.
• cuDNN (> = v3).我们建议使用6.0版.有关详细信息，请参阅NVIDIA文档，尤其是将适当的路径名附加到LD_LIBRARY_PATH环境变量的说明.

• NVIDIA's Cuda Toolkit (>= 7.0). We recommend version 9.0. For details, see NVIDIA's documentation. Ensure that you append the relevant Cuda pathnames to the LD_LIBRARY_PATH environment variable as described in the NVIDIA documentation.
• The NVIDIA drivers associated with NVIDIA's Cuda Toolkit.
• cuDNN (>= v3). We recommend version 6.0. For details, see NVIDIA's documentation, particularly the description of appending the appropriate pathname to your LD_LIBRARY_PATH environment variable.

因此，使用TensorFlow GPU后端，CuDNNLSTM与普通LSTM有何不同?找到可用的TensorFlow GPU后端后，会自动选择CuDNNLSTM并替换普通的LSTM吗?

Therefore, how would a CuDNNLSTM differ in any way from a normal LSTM using a TensorFlow GPU backend? Will CuDNNLSTM be automatically selected and replace the normal LSTM when an available TensorFlow GPU backend is found?

推荐答案

您为什么不亲自尝试一下，看看呢? 以我为例，用LSTM训练模型需要10分钟30秒. 只需将通话从LSTM()切换到CuDNNLSTM()，就花费了不到一分钟的时间.

Why don't you try it out for yourself and see? In my case, training a model with LSTM took 10mins 30seconds. Simply switching the call from LSTM() to CuDNNLSTM() took less than a minute.

我还注意到，切换到CuDNNLSTM()也会大大提高model.evaluate()和model.predict()的速度.

I also noticed that switching to CuDNNLSTM() speeds up model.evaluate() and model.predict() substantially as well.

这篇关于Keras中的CuDNNLSTM和LSTM有什么区别?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！