Vowpal Wabbit is a machine learning system with techniques such as online, hashing, allreduce, reductions, learning2search, active, and interactive learning.

LinearOperator is a PyTorch package for abstracting away the linear algebra routines needed for structured matrices (or operators).

CTranslate2 is a C++ and Python library for efficient inference with Transformer models.

The project implements a custom runtime that applies many performance optimization techniques such as weights quantization, layers fusion, batch reordering, etc., to accelerate and reduce the memory usage of Transformer models on CPU and GPU.

Autograd can automatically differentiate native Python and NumPy code. It can handle a large subset of Python's features, including loops, ifs, recursion and closures, and it can even take derivatives of derivatives of derivatives. It supports reverse-mode differentiation (a.k.a. backpropagation), which means it can efficiently take gradients of scalar-valued functions with respect to array-valued arguments, as well as forward-mode differentiation, and the two can be composed arbitrarily. The main intended application of Autograd is gradient-based optimization.

This package provides a functional take on deep learning, compatible with your favorite libraries.

Scikit-learn provides simple and efficient tools for data mining and data analysis.

LibreTranslate is a machine translation API and web interface, powered by the Argos Translate library.

QNNPACK is a library for low-precision neural network inference. It contains the implementation of common neural network operators on quantized 8-bit tensors.

Captum is a model interpretability and understanding library for PyTorch. Captum contains general purpose implementations of integrated gradients, saliency maps, smoothgrad, vargrad and others for PyTorch models. It has quick integration for models built with domain-specific libraries such as torchvision, torchtext, and others.

PyNNDescent provides a Python implementation of Nearest Neighbor Descent for k-neighbor-graph construction and approximate nearest neighbor search.

FANN is a neural network library, which implements multilayer artificial neural networks in C with support for both fully connected and sparsely connected networks.

This package implements a variety of persistent homology algorithms. It provides an interface for

• computing persistence cohomology of sparse and dense data sets

• visualizing persistence diagrams

• computing lowerstar filtrations on images

• computing representative cochains

The MCL algorithm is short for the Markov Cluster Algorithm, a fast and scalable unsupervised cluster algorithm for graphs (also known as networks) based on simulation of (stochastic) flow in graphs.

This package provides logging utilities for the SpaCy natural language processing framework.

LightGBM is a gradient boosting framework that uses tree based learning algorithms. It is designed to be distributed and efficient with the following advantages:

• Faster training speed and higher efficiency

• Lower memory usage

• Better accuracy

• Parallel and GPU learning supported (not enabled in this package)

• Capable of handling large-scale data

GPy is a Gaussian Process (GP) framework written in Python, from the Sheffield machine learning group. GPy implements a range of machine learning algorithms based on GPs.

OpenFst is a library for constructing, combining, optimizing, and searching weighted finite-state transducers (FSTs).

LIBSVM is a machine learning library for support vector classification, (C-SVC, nu-SVC), regression (epsilon-SVR, nu-SVR) and distribution estimation (one-class SVM). It supports multi-class classification.

This tool provides ordinary differential equation solvers implemented in PyTorch. Backpropagation through ODE solutions is supported using the adjoint method for constant memory cost.

This package provides a tool for visualizing live, rich data for Torch and Numpy.

Lap is a linear assignment problem solver using Jonker-Volgenant algorithm for dense (LAPJV) or sparse (LAPMOD) matrices.

Magic-Wormhole is a library and a command-line tool named wormhole, which makes it possible to securely transfer arbitrary-sized files and directories (or short pieces of text) from one computer to another. The two endpoints are identified by using identical "wormhole codes": in general, the sending machine generates and displays the code, which must then be typed into the receiving machine.

The codes are short and human-pronounceable, using a phonetically-distinct wordlist. The receiving side offers tab-completion on the codewords, so usually only a few characters must be typed. Wormhole codes are single-use and do not need to be memorized.

This package provides the Magic-Wormhole Transit Relay server, which helps clients establish bulk-data transit connections even when both are behind NAT boxes. Each side makes a TCP connection to this server and presents a handshake. Two connections with identical handshakes are glued together, allowing them to pretend they have a direct connection.

This package provides the main server that Magic-Wormhole clients connect to. The server performs store-and-forward delivery for small key-exchange and control messages. Bulk data is sent over a direct TCP connection, or through a transit-relay.