MUSE: Multilingual Unsupervised and Supervised Embeddings

MUSE is a Python library for multilingual word embeddings.

NextSegmPred

NextSegmPred reproduces the results obtained with the S2S (segmentation to segmentation) model described in the paper. Frames with no border correspond to the input while red borders indicate predicted frames.

CLEVR Dataset Generator

This is the code used to generate the CLEVR dataset as described in the paper:  CLEVR: A Diagnostic Dataset for Compositional Language and Elementary Visual Reasoning. Presented at CVPR 2017

StarSpace

StarSpace is a general-purpose neural model for efficient learning of entity embeddings for solving a wide variety of problems.

Noise-as-Targets

This is the code used for unsupervised training of convolutional neural networks as described in the ICML 2017 paper Unsupervised Learning by Predicting Noise.

ParlAI

ParlAI, is a unified platform, implemented in Python, for training and evaluating AI models on a variety of openly available dialog datasets using open-sourced learning agents.

Supervised-Hashing Baselines

This repository contains code to reproduce the baselines in How should we evaluate supervised hashing?

VoiceLoop

VoiceLoop is a neural text-to-speech (TTS) that is able to transform text to speech in voices that are sampled in the wild.

SparseConvNet

SparseConvNet implements spatially sparse convolutional networks.

DrQA

This is a PyTorch implementation of the DrQA system.

ELF

ELF is an Extensive, Lightweight and Flexible platform for game research, in particular for real-time strategy (RTS) games.

InferSent 

InferSent is a sentence embeddings method that provides semantic sentence representations.

SentEval

SentEval is a library for evaluating the quality of sentence embeddings.

End-to-End Negotiator

This is a PyTorch implementation of research paper Deal or No Deal? End-to-End Learning for Negotiation Dialogues developed by Facebook AI Research.

Gradient Episodic Memory for Continual Learning

One major obstacle towards AI is the poor ability of models to solve new problems quicker, and without forgetting previously acquired knowledge. To better understand this issue, we study the problem of continual learning, where the model observes, once and one by one, examples concerning a sequence of tasks.

Inferring and Executing Programs for Visual Reasoning (clevr-iep)

Existing methods for visual reasoning attempt to directly map inputs to outputs using black-box architectures without explicitly modeling the underlying reasoning processes. As a result, these black-box models often learn to exploit biases in the data rather than learning to perform visual reasoning.

Facebook AI Research Sequence-to-Sequence Toolkit

The FAIR Sequence-to-Sequence toolkit implements a fully convolutional model for text generation.

CommAI

The CommAI project aims at developing new data-sets and algorithms to develop and evaluate general-purpose artificial agents that rely on a linguistic interface, and are capable of quickly adapting to a never-ending stream of tasks.

Visdom

When you perform scientific experiments on remote servers, it can be a hassle to produce live visualizations of these experiments and to keep those visualizations organized. Visdom solves this problem for you!

Faiss

Faiss is a library for efficient similarity search and clustering of dense vectors. It contains algorithms that search in sets of vectors of any size, up to ones that possibly do not fit in RAM.