In January this year, Keras author and Google AI researcher Francois Chollet sent out a call on Twitter: Chinese speaking Keras users, would anyone like to help build a Chinese version of the Keras documentation?

More than a month later, the official Chinese document arrived.

Keras is a framework developed by Francois Chollet, an engineer at Google. Running with TensorFlow, CNTK, or Theano as a back end, Keras helps you quickly build and train your own deep learning models. Keras is one of the fastest growing deep learning frameworks. In order to help developers understand and master Keras, we have prepared the following content about Keras for you.


Official resources: ๐Ÿ’ผ

  1. ๐Ÿ‘จ website: keras.io
  2. ๐Ÿ“– Chinese version: keras. IO /zh/
  3. ๐Ÿƒ Quick Start: keras. IO /zh/#30-kera…
  4. ๐Ÿ™ Github:github.com/keras-team/…
  5. ๐Ÿ—ฃ Google+ Online forum: groups.google.com/forum/#! The for…
  6. ๐Ÿฅ slack๏ผškerasteam.slack.com/

The article ๐Ÿ“ƒ

  1. “Hello World “for Neural Networks — An Introduction to Keras

    Francois Chollet, the inventor of the Keras framework, recommended the translation of the article, given the tutorial due to the similarity between Keras and Scikit-Learn: using Keras by comparison with Scikit-Learn.

  2. Five steps to build a neural network

    Creating and evaluating deep neural networks using Keras is very convenient, but you need to follow a few strict steps to build the model. In this article, we will explore the creation, training, and evaluation of deep neural networks in Keras step by step, and learn how to use the trained model for prediction.

  3. Keras TensorFlow: How to develop a complex Deep Learning model from zero

    This article assumes that you are already familiar with TensorFlow and convolutional neural networks. If you are not, check out this 10-minute introduction to TensorFlow and convolutional neural networks tutorial and then come back to this article.

  4. Controlling deep learning training with wechat: Keras plug-in with Chinese characteristics

    Inspired by wechat monitoring Tensorflow training, the author developed a plug-in to control Keras training (control PyTorch, Mxnet, caffe to you, dynamic map class may also be defined with wechat). Now it has realized many functions including passive monitoring, active query, remote shutdown/stop training and so on.

    Tutorial ๐Ÿž

    Here are some suggestions from Keras inventor Francois Chollet:

    If you are not familiar with deep learning and machine learning, you may want to make sure that you have followed the following tutorial. As long as you have some Python background, these basic courses are easy to follow

    • A video tutorial on neural networks and Keras from scratch (CDS.cern. ch/record/2157…)
    • Keras “Hello World” from FastForwardLabs (github.com/fastforward…)
    • “By Keras, step by step to use Python develop your first neural network” (machinelearningmastery.com/tutorial-fi…).

    If you already know something about machine learning and deep learning, here’s the quickest way to get started:

    • Read Keras README (github.com/fchollet/ke…)
    • Reading sequence Models (keras.io/getting-sta…)
    • Reading API (keras.io/getting-sta…)

    Read some key Keras code:

    • MLP (github.com/fchollet/ke…).
    • Convnet (github.com/fchollet/ke…).
    • LSTM (github.com/fchollet/ke…).

    Read the tutorial on Keras’s blog:

    • Building an image classifier with a small amount of data (blog.keras. IO/builder-po…)
    • Using pre-trained word embedding (blog.keras. IO /using-pre-t…)
    • Build an autoencoder using Keras (blog.keras. IO/build-au…) (blog. Keras. IO/building – au…
    • Then apply the skills you’ve learned to real-world problems by participating in a Kaggle competition. Francois Chollet-Session on Aug 15, 2016 – Quora Quora (github.com/fchollet/ke…) You can find many tutorials and code examples here.