Search Results for "reinforcement learning" - Page 4
Sort By:
Showing 211 open source projects for "reinforcement learning"
View related business solutions-
Forever Free Full-Stack Observability | Grafana CloudBuilt on open standards like Prometheus and OpenTelemetry, Grafana Cloud includes Kubernetes Monitoring, Application Observability, Incident Response, plus the AI-powered Grafana Assistant. Get started with our generous free tier today.
-
Enterprise-grade ITSM, for every businessFreshservice is an intuitive, AI-powered platform that helps IT, operations, and business teams deliver exceptional service without the usual complexity. Automate repetitive tasks, resolve issues faster, and provide seamless support across the organization. From managing incidents and assets to driving smarter decisions, Freshservice makes it easy to stay efficient and scale with confidence.
-
1
PilottAI
Python framework for building scalable multi-agent systems
pilottai is an AI-based autonomous drone navigation system utilizing reinforcement learning for real-time decision-making. It is designed for simulating and training drones to fly safely through dynamic environments using AI-based controllers. -
2
AgentEvolver
Towards Efficient Self-Evolving Agent System
...The system focuses on improving the efficiency and scalability of training autonomous agents by allowing them to generate tasks, explore environments, and refine strategies without heavy reliance on manually curated datasets. Its architecture combines reinforcement learning with LLM-driven reasoning mechanisms to guide exploration and learning. The framework introduces several key mechanisms, including self-questioning to create new learning tasks, self-navigating to improve exploration through experience reuse, and self-attributing to assign rewards based on the usefulness of actions. ... -
3
PyBoy
Game Boy emulator written in Python
PyBoy is an open-source Game Boy emulator written in Python, designed for both gameplay and AI experimentation. It allows users to run classic Game Boy games while providing a powerful API for automation, scripting, and reinforcement learning. Developers can interact directly with game memory, inputs, and screen data, making it ideal for training bots and analyzing game mechanics. PyBoy emphasizes performance, enabling accelerated emulation speeds and frame skipping for large-scale simulations. It integrates with tools like OpenAI Gym, allowing seamless use in machine learning workflows. ... -
4
Unsloth Studio
Unified web UI for training and running open models locally
...Built on top of the Unsloth framework, it focuses on high-performance training with reduced VRAM usage and faster speeds compared to traditional methods. The platform supports fine-tuning, pretraining, and reinforcement learning workflows, making it suitable for both experimentation and production use. Users can interact with models through chat, upload files like PDFs or images, and execute code within the environment to improve outputs. By combining powerful optimization techniques with an intuitive UI, Unsloth Studio simplifies the process of building and customizing AI models locally. -
Go From AI Idea to AI App FastAccess Gemini 3 and 200+ models. Build chatbots, agents, or custom models with built-in monitoring and scaling.
-
5
Pearl
A Production-ready Reinforcement Learning AI Agent Library
Pearl is a production-ready reinforcement learning and contextual bandit agent library built for real-world sequential decision making. It is organized around modular components—policy learners, replay buffers, exploration strategies, safety modules, and history summarizers—that snap together to form reliable agents with clear boundaries and strong defaults. The library implements classic and modern algorithms across two regimes: contextual bandits (e.g., LinUCB, LinTS, SquareCB, neural bandits) and fully sequential RL (e.g., DQN, PPO-style policy optimization), with attention to practical concerns like nonstationarity and dynamic action spaces. ... -
6
MiniOneRec
Minimal reproduction of OneRec
...The framework provides an end-to-end pipeline for building generative recommender systems, including semantic identifier construction, supervised fine-tuning, and reinforcement learning-based optimization. Semantic IDs are created using techniques such as quantized variational autoencoders to convert item features into token sequences that can be modeled by transformer architectures. Developers can train and evaluate recommendation models using different backbone language models while benefiting from the generative framework’s parameter efficiency and scalability. -
7
AgentScope
Build and run agents you can see, understand and trust
...With built-in support for ReAct agents, memory, planning, human-in-the-loop control, and real-time voice interaction, developers can create powerful agents in minutes. AgentScope integrates seamlessly with tools, long-term memory systems, MCP, A2A (Agent-to-Agent) protocols, and observability frameworks. It also supports reinforcement learning workflows for tuning agents and improving performance across complex tasks. Deployable locally, serverless in the cloud, or on Kubernetes with OpenTelemetry support, AgentScope is built for both experimentation and production environments. -
8
MuJoCo Playground
An open source library for GPU-accelerated robot learning
...MuJoCo Playground supports both the MJX JAX implementation and the Warp physics engine, enabling flexible use across research pipelines. The environments are designed for fast training, compatibility with reinforcement learning libraries, and real-time trajectory visualization using rscope. -
9
Newton
An open-source, GPU-accelerated physics simulation engine
Newton is a high-performance, GPU-accelerated physics simulation engine designed primarily for robotics research, machine learning, and advanced simulation workflows. Built on top of NVIDIA Warp, it leverages GPU parallelism to deliver scalable and efficient simulation environments that support rapid iteration and experimentation. The engine extends previous simulation frameworks by introducing differentiable physics capabilities, allowing it to integrate seamlessly with machine learning... -
MongoDB Atlas runs apps anywhereMongoDB Atlas gives you the freedom to build and run modern applications anywhere—across AWS, Azure, and Google Cloud. With global availability in over 115 regions, Atlas lets you deploy close to your users, meet compliance needs, and scale with confidence across any geography.
-
10
DeepTutor
AI-Powered Personalized Learning Assistant
DeepTutor is an AI-powered tutoring and learning assistant framework designed to automatically teach, explain, and reinforce academic or technical concepts in depth according to a learner’s specific needs. It goes beyond simple Q&A by constructing multi-stage educational narratives, breaking down complex topics into sequenced “lesson steps,” and offering prompts, examples, and exercises that build on each other in a logical curriculum. The core architecture combines LLM-based reasoning with... -
11
Agent Behavior Monitoring
The open source post-building layer for agents
...Developers can use the framework to observe agent actions in both online production environments and offline evaluation settings, making it useful for debugging and performance analysis. Judgeval transforms agent interaction trajectories into structured evaluation datasets that can be used for reinforcement learning, supervised fine-tuning, or other forms of post-training improvement. The framework includes tools that analyze agent behavior patterns and group interaction trajectories by behavior type or topic, allowing researchers to detect weaknesses or unexpected behaviors. -
12
LLM Colosseum
Benchmark LLMs by fighting in Street Fighter 3
...This setup creates a dynamic environment that tests reasoning, situational awareness, and decision-making abilities in real time. Instead of relying purely on reward signals as in reinforcement learning agents, the models analyze contextual information and generate strategic actions based on the game environment. Performance is evaluated using a competitive ranking system that assigns models an ELO rating based on their results across matches against other models. -
13
Xtuner
A Next-Generation Training Engine Built for Ultra-Large MoE Models
...Its architecture incorporates memory-efficient optimizations that allow researchers to train large models even when computational resources are limited. XTuner is also designed to integrate with modern AI ecosystems, supporting multimodal training, reinforcement learning optimization, and instruction tuning pipelines. -
14
slime LLM
slime is an LLM post-training framework for RL Scaling
slime is an open-source large language model (LLM) post-training framework developed to support reinforcement learning (RL)-based scaling and high-performance training workflows for advanced LLMs, blending training and rollout modules into an extensible system. It offers a flexible architecture that connects high-throughput training (e.g., via Megatron-LM) with a customizable data generation pipeline, enabling researchers and engineers to iterate on new RL training paradigms effectively. ... -
15
Agent Lightning
The absolute trainer to light up AI agents
Agent Lightning is an open-source framework developed by Microsoft to train and optimize AI agents using techniques like reinforcement learning (RL), supervised fine-tuning, and automatic prompt optimization, with minimal or zero changes to existing agent code. It’s designed to be compatible with a wide range of agent architectures and frameworks — from LangChain and OpenAI Agent SDKs to AutoGen and custom Python agents — making it broadly applicable across different agent tooling ecosystems. ... -
16
NVIDIA NeMo Agent Toolkit
Library for efficiently connecting and optimizing teams of AI agents
...Developers can monitor agent execution, trace workflows, and analyze token-level performance to identify bottlenecks and improve efficiency. NeMo Agent Toolkit also supports evaluation systems, prompt optimization, and reinforcement learning techniques to enhance agent behavior over time. By combining instrumentation, workflow orchestration, and performance optimization tools, the platform helps developers deploy scalable and intelligent multi-agent systems. -
17
GLM-TTS
Controllable & emotion-expressive zero-shot TTS
...It uses a two-stage architecture where a generative LLM first converts text into intermediate speech token sequences and then a Flow-based neural model converts those tokens into natural audio waveforms, enabling rich prosody and voice character even for unseen speakers. The system introduces a multi-reward reinforcement learning framework that jointly optimizes for voice similarity, emotional expressiveness, pronunciation, and intelligibility, yielding output that can rival commercial options in naturalness and expressiveness. GLM-TTS also supports phoneme-level control and hybrid text + phoneme input, giving developers precise control over pronunciation critical for multilingual or polyphone-rich languages. -
18
SWIFT LLM
Use PEFT or Full-parameter to CPT/SFT/DPO/GRPO 600+ LLMs
SWIFT LLM is a comprehensive framework developed within the ModelScope ecosystem for training, fine-tuning, evaluating, and deploying large language models and multimodal models. The platform provides a full machine learning pipeline that supports tasks ranging from model pre-training to reinforcement learning alignment techniques. It integrates with popular inference engines such as vLLM and LMDeploy to accelerate deployment and runtime performance. The framework also includes support for many modern training strategies, including preference learning methods and parameter-efficient fine-tuning techniques. ms-swift is designed to work with hundreds of language and multimodal models, providing a unified environment for experimentation and production deployment. -
19
ML for Beginners
12 weeks, 26 lessons, 52 quizzes, classic Machine Learning for all
ML-For-Beginners is a structured, project-driven curriculum that teaches foundational machine learning concepts with approachable math and lots of code. Organized as a multi-week course, it mixes short lectures with labs in notebooks so learners practice regression, classification, clustering, and recommendation techniques on real datasets. Each lesson aims to connect the algorithm to a relatable scenario, reinforcing intuition before diving into parameters, metrics, and trade-offs. The... -
20
GLM-4
GLM-4 series: Open Multilingual Multimodal Chat LMs
...The GLM-4-32B-0414 models are trained on ~15T high-quality data (including substantial synthetic reasoning data), then post-trained with preference alignment, rejection sampling, and reinforcement learning to improve instruction following, coding, function calling, and agent-style behaviors. The GLM-Z1-32B-0414 line adds deeper mathematical, coding, and logical reasoning via extended reinforcement learning and pairwise ranking feedback, while GLM-Z1-Rumination-32B-0414 introduces a “rumination” mode that performs longer, tool-using deep research for complex, open-ended tasks. ... -
21
crème de la crème of AI courses
This repository is a curated collection of links to various courses
...Topics covered include deep learning, natural language processing, computer vision, large language models, linear algebra, reinforcement learning, and machine learning engineering. Because the repository links to well-known educational content such as university lecture series and professional training materials, it functions as a structured roadmap for individuals who want to develop expertise in artificial intelligence. -
22
LatentMAS
Latent Collaboration in Multi-Agent Systems
LatentMAS is an advanced framework for multi-agent reinforcement learning (MARL) that uses latent variable modeling to bridge perception and decision-making in environments where agents must coordinate under uncertainty. It provides mechanisms for agents to learn high-level latent representations of states, which simplifies complex sensory inputs into compact, actionable embeddings that facilitate both individual policy learning and inter-agent coordination. -
23
GLM-4.5V
GLM-4.6V/4.5V/4.1V-Thinking, towards versatile multimodal reasoning
...It embodies the design philosophy of mixing visual and textual modalities into a unified model capable of general-purpose reasoning, content understanding, and generation, while already supporting a wide variety of tasks: from image captioning and visual question answering to content recognition, GUI-based agents, video understanding, and long-document interpretation. GLM-4.5V emerged from a training framework that leverages scalable reinforcement learning (with curriculum sampling) to boost performance across tasks ranging from STEM problem solving to long-context reasoning, giving it broad applicability beyond narrow benchmarks. When it was released, it achieved state-of-the-art results on a large collection of public multimodal benchmarks for open-source models. -
24
Diplomacy Cicero
Code for Cicero, an AI agent that plays the game of Diplomacy
...It is designed to play the board game Diplomacy by combining open-domain natural language negotiation with strategic planning. The repository includes training code, model checkpoints, and infrastructure for both language modelling (via the ParlAI framework) and reinforcement learning for strategy agents. It supports two variants: Cicero (which handles full “press” negotiation) and Diplodocus (a variant focused on no-press diplomacy) as described in the README. The codebase is implemented primarily in Python with performance-critical components in C++ (via pybind11 bindings) and is configured to run in a high‐GPU cluster environment. ... -
25
Agent S
Agent S: an open agentic framework that uses computers like a human
Agent S is an open-source agentic framework designed to enable autonomous computer use through an Agent-Computer Interface (ACI). Built to operate graphical user interfaces like a human, it allows AI agents to perceive screens, reason about tasks, and execute actions across macOS, Windows, and Linux systems. The latest version, Agent S3, surpasses human-level performance on the OSWorld benchmark, demonstrating state-of-the-art results in complex multi-step computer tasks. Agent S combines...
