Libraries, frameworks and developer tools for building, training, running and connecting AI models.
62 terms
AI Fairness 360 (AIF360) is an open-source toolkit that measures bias in datasets and models and applies algorithms to mitigate it in machine learning.
The Apple Foundation Models framework is a Swift API that lets apps run the on-device Apple Intelligence language model privately, offline and at no cost.
AutoGen is an open-source Microsoft framework for building multi-agent applications in which language model agents converse, use tools and solve tasks together.
CatBoost is an open-source gradient boosting library that handles categorical features natively and resists overfitting through ordered boosting.
The Claude Agent SDK is a Python and TypeScript library that exposes the agent loop behind Claude Code for building autonomous, tool-using AI agents.
ComfyUI is an open-source, node-based interface and execution engine for building, running and sharing diffusion model image and video workflows.
Core ML is the Apple framework that runs trained machine learning models on iPhone, iPad, Mac, Apple Watch and Vision Pro devices without a server.
CrewAI is an open-source Python framework for building teams of role-based AI agents that share tasks, tools and context to complete multi-step work.
DeepSpeed is an open-source PyTorch library that partitions memory and parallelizes work so very large neural networks can be trained and served efficiently.
Diffusers is the open-source Python AI library from Hugging Face for running, fine-tuning and training diffusion models that generate images, video and audio.
DSPy is an open-source Python framework that builds language model pipelines from declarative modules and optimizes their prompts and weights against a metric.
ExecuTorch is the PyTorch runtime for running exported AI models on phones, wearables, microcontrollers and other edge devices without Python.
Fairlearn is an open-source Python toolkit for measuring group fairness in machine learning models and reducing disparities with mitigation algorithms.
fastai is an open-source deep learning library built on PyTorch that offers a layered API for training accurate models with little code and sound defaults.
Flax is an open-source neural network library for JAX that adds modules, parameter handling and training utilities on top of composable transformations.
Google Colab is a hosted Jupyter notebook service that runs Python code on cloud virtual machines with optional GPU and TPU accelerators for machine learning.
Gradio is an open-source Python library that wraps machine learning models and functions in interactive web interfaces and APIs with a few lines of code.
Gymnasium is the open-source Python library that defines a standard API and reference environments for training and testing reinforcement learning agents.
Haystack is an open-source Python framework from deepset for building retrieval-augmented generation, search and agent pipelines on language models.
The Hugging Face Hub is a Git-versioned platform for hosting and sharing machine learning models, datasets and demo apps, publicly or privately.
Hugging Face Transformers is an open-source Python library that loads, fine-tunes and runs pretrained transformer models through one consistent interface.
JAX is a Python library for array computing that combines a NumPy-style API with automatic differentiation, XLA compilation and parallel execution.
Kaggle is a data science platform for machine learning competitions, public datasets, hosted notebooks with free GPUs and shared pretrained models.
Keras is a high-level deep learning API for Python that defines, trains and deploys neural networks on JAX, TensorFlow or PyTorch backends.
LangChain is an open-source framework for building applications on large language models by composing models, prompts, retrievers and tools.
Langfuse is an open source LLM engineering platform for tracing, prompt management and evaluation of applications built on large language models.
LangGraph is an open-source library for building stateful, controllable AI agents and multi-step workflows as graphs of nodes, edges and shared state.
LangSmith is a platform from LangChain for tracing, evaluating and monitoring applications built on large language models and AI agents.
LightGBM is an open-source gradient boosting framework that trains fast, accurate decision tree ensembles on large tabular datasets with low memory use.
LiteLLM is an open-source Python library and proxy that gives applications one OpenAI-compatible interface to more than a hundred language model providers.
LiteRT is Google's on-device runtime, formerly TensorFlow Lite, that runs compact machine learning models efficiently on phones, embedded boards and browsers.
llama.cpp is an open-source C/C++ engine that runs quantized large language models locally on CPUs and GPUs using the GGUF file format.
LlamaIndex is an open-source data framework that connects large language models to private documents through ingestion, indexing and retrieval pipelines.
LM Studio is a desktop application for downloading, running and serving open-weight language models locally on a personal computer.
MediaPipe is an open-source Google framework for building on-device machine learning pipelines that process live video, images, audio and text in real time.
MLflow is an open-source platform for tracking machine learning experiments, packaging models, managing model versions and tracing generative AI applications.
NLTK, the Natural Language Toolkit, is an open-source Python library for classic text processing, linguistic analysis and teaching natural language processing.
NNAPI, the Android Neural Networks API, is a C interface that lets apps run neural network inference on phone GPUs, DSPs and NPUs.
Ollama is an open-source tool that downloads, manages and serves large language models on a local computer through a simple CLI and REST API.
ONNX (Open Neural Network Exchange) is an open file format that stores trained machine learning models as portable computation graphs for many runtimes.
ONNX Runtime is a cross-platform inference engine that runs ONNX models on CPUs, GPUs and NPUs through pluggable hardware execution providers.
Open WebUI is a self-hosted, open-source web interface for chatting with local and remote language models, with users, documents and tools in one place.
The OpenAI Agents SDK is an open-source framework for building AI agents from instructions, tools, handoffs, guardrails and built-in tracing.
Pydantic AI is a Python agent framework that builds type-safe LLM applications with validated structured output, typed tools and dependency injection.
PyTorch is an open-source deep learning framework that combines Python-first tensor computation, automatic differentiation and GPU acceleration.
PyTorch Lightning is an open-source framework that organizes PyTorch code and automates training loops, hardware scaling, logging and checkpointing.
Ray is an open-source distributed computing framework that scales Python and AI workloads from a laptop to a cluster with tasks, actors and an object store.
safetensors is a simple, secure file format for storing model weights that loads tensors without running code and supports fast memory-mapped access.
scikit-learn is an open-source Python library for classical machine learning, offering a uniform estimator API for modeling, preprocessing and evaluation.
Semantic Kernel is an open-source Microsoft SDK that connects large language models to application code through plugins, connectors and function calling.