This folder contains machine learning models implemented by researchers in TensorFlow. The models are maintained by their respective authors. To propose a model for inclusion, please submit a pull request.

• adversarial_crypto: protecting communications with adversarial neural cryptography.
• adversarial_text: semi-supervised sequence learning with adversarial training.
• attention_ocr: a model for real-world image text extraction.
• audioset: Models and supporting code for use with AudioSet.
• autoencoder: various autoencoders.
• brain_coder: Program synthesis with reinforcement learning.
• cognitive_mapping_and_planning: implementation of a spatial memory based mapping and planning architecture for visual navigation.
• compression: compressing and decompressing images using a pre-trained Residual GRU network.
• cvt_text: semi-supervised sequence learning with cross-view training.
• deep_contextual_bandits: code for a variety of contextual bandits algorithms using deep neural networks and Thompson sampling.
• deep_speech: automatic speech recognition.
• deeplab: deep labeling for semantic image segmentation.
• delf: deep local features for image matching and retrieval.
• differential_privacy: differential privacy for training data.
• domain_adaptation: domain separation networks.
• fivo: filtering variational objectives for training generative sequence models.
• gan: generative adversarial networks.
• im2txt: image-to-text neural network for image captioning.
• inception: deep convolutional networks for computer vision.
• keypointnet: discovery of latent 3D keypoints via end-to-end geometric eeasoning [demo].
• learning_to_remember_rare_events: a large-scale life-long memory module for use in deep learning.
• learning_unsupervised_learning: a meta-learned unsupervised learning update rule.
• lexnet_nc: a distributed model for noun compound relationship classification.
• lfads: sequential variational autoencoder for analyzing neuroscience data.
• lm_1b: language modeling on the one billion word benchmark.
• lm_commonsense: commonsense reasoning using language models.
• maskgan: text generation with GANs.
• namignizer: recognize and generate names.
• neural_gpu: highly parallel neural computer.
• neural_programmer: neural network augmented with logic and mathematic operations.
• next_frame_prediction: probabilistic future frame synthesis via cross convolutional networks.
• object_detection: localizing and identifying multiple objects in a single image.
• pcl_rl: code for several reinforcement learning algorithms, including Path Consistency Learning.
• ptn: perspective transformer nets for 3D object reconstruction.
• marco: automating the evaluation of crystallization experiments.
• qa_kg: module networks for question answering on knowledge graphs.
• real_nvp: density estimation using real-valued non-volume preserving (real NVP) transformations.
• rebar: low-variance, unbiased gradient estimates for discrete latent variable models.
• resnet: deep and wide residual networks.
• seq2species: deep learning solution for read-level taxonomic classification.
• skip_thoughts: recurrent neural network sentence-to-vector encoder.
• slim: image classification models in TF-Slim.
• street: identify the name of a street (in France) from an image using a Deep RNN.
• struct2depth: unsupervised learning of depth and ego-motion.
• swivel: the Swivel algorithm for generating word embeddings.
• syntaxnet: neural models of natural language syntax.
• tcn: Self-supervised representation learning from multi-view video.
• textsum: sequence-to-sequence with attention model for text summarization.
• transformer: spatial transformer network, which allows the spatial manipulation of data within the network.
• vid2depth: learning depth and ego-motion unsupervised from raw monocular video.
• video_prediction: predicting future video frames with neural advection.