Open Source Python Artificial Intelligence Software

Fast Artificial Neural Network Library is a free open source neural network library, which implements multilayer artificial neural networks in C with support for both fully connected and sparsely connected networks. Cross-platform execution in both fixed and floating point are supported. It includes a framework for easy handling of training data sets. It is easy to use, versatile, well documented, and fast. Bindings to more than 15 programming languages are available. An easy to read introduction article and a reference manual accompanies the library with examples and recommendations on how to use the library. Several graphical user interfaces are also available for the library.

fairseq2 is a modern, modular sequence modeling framework developed by Meta AI Research as a complete redesign of the original fairseq library. Built from the ground up for scalability, composability, and research flexibility, fairseq2 supports a broad range of language, speech, and multimodal content generation tasks, including instruction fine-tuning, reinforcement learning from human feedback (RLHF), and large-scale multilingual modeling. Unlike the original fairseq—which evolved into a large, monolithic codebase—fairseq2 introduces a clean, plugin-oriented architecture designed for long-term maintainability and rapid experimentation. It supports multi-GPU and multi-node distributed training using DDP, FSDP, and tensor parallelism, capable of scaling up to 70B+ parameter models. The framework integrates seamlessly with PyTorch 2.x features such as torch.compile, Fully Sharded Data Parallel (FSDP), and modern configuration management.

Kaldi is an open source toolkit for speech recognition research. It provides a powerful framework for building state-of-the-art automatic speech recognition (ASR) systems, with support for deep neural networks, Gaussian mixture models, hidden Markov models, and other advanced techniques. The toolkit is widely used in both academia and industry due to its flexibility, extensibility, and strong community support. Kaldi is designed for researchers who need a highly customizable environment to experiment with new algorithms, as well as for practitioners who want robust, production-ready ASR pipelines. It includes extensive tools for data preparation, feature extraction, acoustic and language modeling, decoding, and evaluation. With its modular design, Kaldi allows users to adapt the system to a wide range of languages and domains. As one of the most influential projects in speech recognition, it has become a foundation for much of the modern work in ASR.

The Player Project: Player is a networked interface to robots and sensors. Stage and Gazebo are Player-friendly multiple-robot simulators. The software aims for POSIX compliance and runs on most UNIX-like OS's. Some parts also work on Windows.

llama2.c is a minimalist implementation of the Llama 2 language model architecture designed to run entirely in pure C. Created by Andrej Karpathy, this project offers an educational and lightweight framework for performing inference on small Llama 2 models without external dependencies. It provides a full training and inference pipeline: models can be trained in PyTorch and later executed using a concise 700-line C program (run.c). While it can technically load Meta’s official Llama 2 models, current support is limited to fp32 precision, meaning practical use is capped at models up to around 7B parameters. The goal of llama2.c is to demonstrate how a compact and transparent implementation can perform meaningful inference even with small models, emphasizing simplicity, clarity, and accessibility. The project builds upon lessons from nanoGPT and takes inspiration from llama.cpp, focusing instead on minimalism and educational value over large-scale performance.

Presage (formerly Soothsayer) is an intelligent predictive text entry system. Presage generates predictions by modelling natural language as a combination of redundant information sources. Presage computes probabilities for words which are most likely to be entered next by merging predictions generated by the different predictive algorithms. Presage's modular and extensible architecture allows its language model to be extended and customized to utilize statistical, syntactic, and semantic predictive algorithms. Presage's predictive capabilities are implemented by predictive plugins. Predictive plugins use services provided by the platform to implement multiple prediction techniques.

mujoco-py is a Python wrapper for MuJoCo, a high-performance physics engine widely used in robotics, reinforcement learning, and AI research. It allows developers and researchers to run detailed rigid body simulations with contacts directly from Python, making MuJoCo easier to integrate into machine learning workflows. The library is compatible with MuJoCo version 2.1 and supports Linux and macOS, while Windows support has been deprecated. It provides utilities for loading models, running simulations, and accessing simulation states in real time, along with visualization tools for rendering environments. The project also includes interactive examples showcasing collision handling, texture randomization, state resetting, and robot control. By bridging MuJoCo with Python, mujoco-py enables rapid prototyping, training, and evaluation of AI agents in physics-rich environments.

Python module to interface the CLIPS expert system shell library.

FramerD is a distributed semi-structured object database originally developed at MIT. It provides an internationalized Scheme-based scripting language, built-in text analysis tools, and special support for web scripting.

The main intention of this project is to support the SNNS (Stuttgart Neural Net Simulator) with patches, bugfixes, tools and add on developments. Second target is to publish useful links and information about the SNNS.

We experiment with Evolution of Artifical Neural Networks, combining the two fields of Evolutionary Computation and ANNs. Our methods are applied to a variety of interesting problems. To learn more, click on "Home Page", "Mail", or "Files".

The General Hidden Markov Model Library (GHMM) is a C library with additional Python bindings implementing a wide range of types of Hidden Markov Models and algorithms: discrete, continous emissions, basic training, HMM clustering, HMM mixtures.

Uranie is a sensitivity and uncertainty analysis plateform based on the ROOT framework (http://root.cern.ch) . It is developed at CEA, the French Atomic Energy Commission (http://www.cea.fr). It provides various tools for: - data analysis - sampling - statistical modeling - optimisation - sensitivity analysis - uncertainty analysis - running code on high performance computers - etc. Thanks to ROOT, it is easily scriptable in CINT (c++ like syntax) and Python. Is is available both for Unix and Windows platforms (a dedicated platform archive is available on request). Note : if you have downloaded version 3.12 before the 8th of february, a patch exists for a minor bug on TOutputFileKey file, don't hesitate to ask us.

mlpy is a Python module for Machine Learning built on top of NumPy/SciPy and of GSL. mlpy provides high-level functions and classes allowing, with few lines of code, the design of rich workflows for classification, regression, clustering and feature selection. mlpy is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License version 3. mlpy is available both for Python >=2.6 and Python 3.X.

Vinux supports software text to speech and Braille support from boot-up to shutdown. Users can use installation medium to install independently with no sighted assistance required. Vinux supports command line environment speech, Desktop environment speech and magnification features. Vinux comes with an accessible suite of software and has an excellent mailing list support group.

NeMo is a high-performance spiking neural network simulator which simulates networks of Izhikevich neurons on CUDA-enabled GPUs. NeMo is a C++ class library, with additional interfaces for pure C, Python, and Matlab.

PROJECT HAS MOVED! ROS is now hosted at https://code.ros.org/gf/project/ros/

Soar is a general cognitive architecture for developing systems that exhibit intelligent behavior. Researchers all over the world, both from the fields of artificial intelligence and cognitive science, are using Soar for a variety of tasks.

TexLexAn is an open source text analyser for Linux, able to estimate the readability and reading time, to classify and summarize texts. It has some learning abilities and accepts html, doc, pdf, ppt, odt and txt documents. Written in C and Python.

THIS PROJECT MIGRATED TO https://gitlab.com/mwetoolkit/mwetoolkit3/ The Multiword Expressions toolkit aids in the automatic identification and extraction of multiword units in running text. These include idioms (kick the bucket), noun compounds (cable car), phrasal verbs (take off, give up), etc. Even though it focuses on multiword expresisons, the framework is quite complete and can also be useful in any corpus-based study in computational linguistics. The mwetoolkit can be applied to virtually any text collection, language, and MWE type. It is a command-line tool written mostly in Python. Its development started in 2010 as a PhD thesis but the project keeps active (see the SVN logs). Up-to-date documentation and details about the tool can be found on the mwetoolkit website: http://mwetoolkit.sourceforge.net/

The Discrete Event Calculus Reasoner is an open source program for performing automated commonsense reasoning using the discrete event calculus, a comprehensive and highly usable formalism for reasoning about action, change, space, and mental states.

Part-of-speech tagging is the task of assigning symbols from a particular set to words in a natural language text. ACOPOST implements and extends well-known machine learning techniques and provides a uniform environment for testing.

The goal of this project is to learn about and develop an AI. Current path is using a bot on AIM.

Betelgeuse is a machine learning modeling package designed to meet the requirements of heavy-duty industry use. It was designed to be efficient, reliable, and highly modular; it is developed primarily in Python to promote maintainability and rapid development, but uses Cython and C in critical bottlenecks for efficiency. It focuses on high-quality implementations of a diverse set of the most widely used machine learning algorithms. An important goal of Betelgeuse is to have a clean, professional user interface amenable to less technical users, and to have multiple user interfaces for graphical, command line, and remote server use.

Brainiac, Is C/C++ Libraries, Programs, And Python, And Lua Scripts For Neural Networking And Genetic Programming, In An Attempt To Create A "Glue-It-All-Together" Project, Striving Towards General Artificial Intelligence