1# Keras examples directory 2 3## Vision models examples 4 5[mnist_mlp.py](mnist_mlp.py) 6Trains a simple deep multi-layer perceptron on the MNIST dataset. 7 8[mnist_cnn.py](mnist_cnn.py) 9Trains a simple convnet on the MNIST dataset. 10 11[cifar10_cnn.py](cifar10_cnn.py) 12Trains a simple deep CNN on the CIFAR10 small images dataset. 13 14[cifar10_cnn_capsule.py](cifar10_cnn_capsule.py) 15Trains a simple CNN-Capsule Network on the CIFAR10 small images dataset. 16 17[cifar10_resnet.py](cifar10_resnet.py) 18Trains a ResNet on the CIFAR10 small images dataset. 19 20[conv_lstm.py](conv_lstm.py) 21Demonstrates the use of a convolutional LSTM network. 22 23[image_ocr.py](image_ocr.py) 24Trains a convolutional stack followed by a recurrent stack and a CTC logloss function to perform optical character recognition (OCR). 25 26[mnist_acgan.py](mnist_acgan.py) 27Implementation of AC-GAN (Auxiliary Classifier GAN) on the MNIST dataset 28 29[mnist_hierarchical_rnn.py](mnist_hierarchical_rnn.py) 30Trains a Hierarchical RNN (HRNN) to classify MNIST digits. 31 32[mnist_siamese.py](mnist_siamese.py) 33Trains a Siamese multi-layer perceptron on pairs of digits from the MNIST dataset. 34 35[mnist_swwae.py](mnist_swwae.py) 36Trains a Stacked What-Where AutoEncoder built on residual blocks on the MNIST dataset. 37 38[mnist_transfer_cnn.py](mnist_transfer_cnn.py) 39Transfer learning toy example on the MNIST dataset. 40 41[mnist_denoising_autoencoder.py](mnist_denoising_autoencoder.py) 42Trains a denoising autoencoder on the MNIST dataset. 43 44---- 45 46## Text & sequences examples 47 48[addition_rnn.py](addition_rnn.py) 49Implementation of sequence to sequence learning for performing addition of two numbers (as strings). 50 51[babi_rnn.py](babi_rnn.py) 52Trains a two-branch recurrent network on the bAbI dataset for reading comprehension. 53 54[babi_memnn.py](babi_memnn.py) 55Trains a memory network on the bAbI dataset for reading comprehension. 56 57[imdb_bidirectional_lstm.py](imdb_bidirectional_lstm.py) 58Trains a Bidirectional LSTM on the IMDB sentiment classification task. 59 60[imdb_cnn.py](imdb_cnn.py) 61Demonstrates the use of Convolution1D for text classification. 62 63[imdb_cnn_lstm.py](imdb_cnn_lstm.py) 64Trains a convolutional stack followed by a recurrent stack network on the IMDB sentiment classification task. 65 66[imdb_fasttext.py](imdb_fasttext.py) 67Trains a FastText model on the IMDB sentiment classification task. 68 69[imdb_lstm.py](imdb_lstm.py) 70Trains an LSTM model on the IMDB sentiment classification task. 71 72[lstm_stateful.py](lstm_stateful.py) 73Demonstrates how to use stateful RNNs to model long sequences efficiently. 74 75[lstm_seq2seq.py](lstm_seq2seq.py) 76Trains a basic character-level sequence-to-sequence model. 77 78[lstm_seq2seq_restore.py](lstm_seq2seq_restore.py) 79Restores a character-level sequence to sequence model from disk (saved by [lstm_seq2seq.py](lstm_seq2seq.py)) and uses it to generate predictions. 80 81[pretrained_word_embeddings.py](pretrained_word_embeddings.py) 82Loads pre-trained word embeddings (GloVe embeddings) into a frozen Keras Embedding layer, and uses it to train a text classification model on the 20 Newsgroup dataset. 83 84[reuters_mlp.py](reuters_mlp.py) 85Trains and evaluate a simple MLP on the Reuters newswire topic classification task. 86 87---- 88 89## Generative models examples 90 91[lstm_text_generation.py](lstm_text_generation.py) 92Generates text from Nietzsche's writings. 93 94[conv_filter_visualization.py](conv_filter_visualization.py) 95Visualization of the filters of VGG16, via gradient ascent in input space. 96 97[deep_dream.py](deep_dream.py) 98Deep Dreams in Keras. 99 100[neural_doodle.py](neural_doodle.py) 101Neural doodle. 102 103[neural_style_transfer.py](neural_style_transfer.py) 104Neural style transfer. 105 106[variational_autoencoder.py](variational_autoencoder.py) 107Demonstrates how to build a variational autoencoder. 108 109[variational_autoencoder_deconv.py](variational_autoencoder_deconv.py) 110Demonstrates how to build a variational autoencoder with Keras using deconvolution layers. 111 112---- 113 114## Examples demonstrating specific Keras functionality 115 116[antirectifier.py](antirectifier.py) 117Demonstrates how to write custom layers for Keras. 118 119[mnist_sklearn_wrapper.py](mnist_sklearn_wrapper.py) 120Demonstrates how to use the sklearn wrapper. 121 122[mnist_irnn.py](mnist_irnn.py) 123Reproduction of the IRNN experiment with pixel-by-pixel sequential MNIST in "A Simple Way to Initialize Recurrent Networks of Rectified Linear Units" by Le et al. 124 125[mnist_net2net.py](mnist_net2net.py) 126Reproduction of the Net2Net experiment with MNIST in "Net2Net: Accelerating Learning via Knowledge Transfer". 127 128[reuters_mlp_relu_vs_selu.py](reuters_mlp_relu_vs_selu.py) 129Compares self-normalizing MLPs with regular MLPs. 130 131[mnist_tfrecord.py](mnist_tfrecord.py) 132MNIST dataset with TFRecords, the standard TensorFlow data format. 133 134[mnist_dataset_api.py](mnist_dataset_api.py) 135MNIST dataset with TensorFlow's Dataset API. 136 137[cifar10_cnn_tfaugment2d.py](cifar10_cnn_tfaugment2d.py) 138Trains a simple deep CNN on the CIFAR10 small images dataset using Tensorflow internal augmentation APIs. 139 140[tensorboard_embeddings_mnist.py](tensorboard_embeddings_mnist.py) 141Trains a simple convnet on the MNIST dataset and embeds test data which can be later visualized using TensorBoard's Embedding Projector.