Agent frameworks & infrastructure
Models, fine-tuning & training
2,504 agents · most starred first
Models, fine-tuning & training, page 7 of 26
Reinforcement learning in Julia. Solving OpenAI gym.
OpenAI LunarLander-v2 DeepRL-based solutions (DQN, DuelingDQN, D3QN)
Multi-Agent Determinantal Q-Learning
A Continual Learning Framework for Production LLM Agents
Generate conversational, tool-calling, structured-output, and preference datasets — easily and at scale
A ridiculously fast Python BPE (Byte Pair Encoder) implementation written in Rust
RLVR training for LLM in CUDA/C++
transformers pytorch safetensors llama text-generation en
transformers safetensors qwen2 text-generation conversational en
transformers pytorch persimmon text-generation license:apache-2.0 endpoints_compatible
transformers pytorch codegen text-generation arxiv:2203.13474 license:bsd-3-clause
transformers safetensors qwen text-generation custom_code zh
Implementation for mSAC methods in PyTorch
Naive Bayes Text Classifier
Eval Protocol (EP) is an open solution for doing reinforcement learning fine-tuning on existing agents — across any language, container, or framework.
transformers safetensors llama text-generation conversational arxiv:2403.04652
transformers safetensors mixtral text-generation merge moe
transformers safetensors llama text-generation conversational arxiv:2403.04652
transformers safetensors qwen2 text-generation chat conversational
Autonomous agentic system for supervised machine learning model development.
transformers safetensors llama text-generation conversational base_model:deepseek-ai/deepseek-coder-33b-instruct
Build conversation Seq2Seq models with TensorFlow
The code implementation of MAGDi: Structured Distillation of Multi-Agent Interaction Graphs Improves Reasoning in Smaller Language Models. Paper: https://arxiv.org/abs/2402.01620
Autonomous ML agent for running experiments using Claude.
[ICLR2026] The official repository for the CodeGym project: "Generalizable End-to-End Tool-Use RL with Synthetic CodeGym"
transformers safetensors llama text-generation facebook meta
transformers safetensors qwen2 text-generation text-generation-inference unsloth
Code for NeurIPS2023 accepted paper: Counterfactual Conservative Q Learning for Offline Multi-agent Reinforcement Learning.
Using deep research workflow to generate datasets for finetuning LLMs.
Important Note fastrl version 2 is being developed at fastrl. Note the link in the readme
The code for paper "EPO: Entropy-regularized Policy Optimization for LLM Agents Reinforcement Learning"
An RL Recipe for Building Agentic LLMs via Self-Imitation on Long-Horizon Agentic Tasks
Compiling strategy guides into reward functions for reinforcement learning. Uses Claude Vision to extract unit tests from game guides, then trains agents with dense, interpretable rewards.
Agentic RL post-training framework
gradio region:us
transformers safetensors llama text-generation code granite
transformers safetensors qwen2 text-generation chat conversational
LatentSkill: From In-Context Textual Skills to In-Weight Latent Skills for LLM Agents
Exploring Multi-Session Agentic Reinforcement Learning
Implementation of Multi-Agent Deep Deterministic Policy Gradients
Official code for "Breaking the Ceiling in On-Policy Distillation via Multi-Agent Debate" (arXiv:2605.01347).
A dependency-free Python genetic-algorithm engine with pluggable fitness criteria — the reusable search core for an LLM/AI autoresearcher.
transformers safetensors qwen3_5 image-text-to-text abliterated uncensored
transformers safetensors mistral text-generation conversational tr
A collection of several Deep Reinforcement Learning techniques (Deep Q Learning, Policy Gradients, ...), gets updated over time.
transformers safetensors llama text-generation code granite
gguf quantized moe vlm vision agentic
transformers gguf llama llama-2 text-generation code
gguf moe expert-pruning rco code agentic
It is a seq2seq encoder decoder chatbot using keras and with attention
An extensible RL framework for training LLM agents with advanced search capabilities, built on VERL and supporting state-of-the-art search strategies.
safetensors qwen2 text-generation conversational tr license:apache-2.0
transformers safetensors llama text-generation conversational arxiv:2403.04652
transformers safetensors qwen2 text-generation code codeqwen
transformers gguf merge moe a3b ties
Agent-Omit: Training Efficient LLM Agents for Adaptive Thought and Observation Omission via Reinforcement Learning
transformers safetensors gpt_bigcode text-generation conversational arxiv:2407.00121
transformers safetensors internlm2 text-generation conversational custom_code
reproduce some RL or Multi-Agent models
无需预算,使用你的个人数据克隆自己——赛博飞升!Clone yourself by tuning a LLM using your own data.
(ICLR 2026 Oral) Code for the paper: Reducing Belief Deviation in Reinforcement Learning for Active Reasoning of LLM Agents
Automated pipeline for generating, verifying, and preparing high-quality function calling datasets for model fine-tuning. Reduces manual annotation through multi-stage validation and conversion to training formats. Built for Azure ML with permissive license models (Phi3.5 MoE, Phi4 etc).
[ICML 2026] This is the official implementation for paper HiPER: Hierarchical Reinforcement Learning with Explicit Credit Assignment for Large Language Model Agents
基于强化学习的 LLM Agent 长链路决策优化系统
gradio region:us
transformers pytorch tensorboard gpt2 text-generation code
transformers gguf llama llama-2 text-generation code
Official codebase for the paper "Sharpening Tax in Post-Training"
transformers safetensors qwen3_next text-generation unsloth qwen
Pytorch Implementation of "Sequential Dialogue Context Modeling for Spoken Language Understanding( https://arxiv.org/abs/1705.03455 )"
Easy-to-follow Pytorch tutorial Notebook for Multi-Agent-Diverse-Generative-Adversarial-Networks
This is a MATLAB-based reinforcement learning framework that includes the Proximal Policy Optimization (PPO) algorithm and its multi-agent extension (MAPPO). It supports GPU acceleration and parallel computing, making it suitable for research and engineering applications in control systems.
AutoAdapt: An Automated Domain Adaptation Framework for Large Language Models
transformers safetensors stablelm text-generation causal-lm conversational
transformers pytorch safetensors llama text-generation conversational
transformers gguf mistral code text-generation-inference text-generation
A pytorch implementation for the MMI-anti model
the small distributed language model toolkit; fine-tune state-of-the-art LLMs anywhere, rapidly
transformers safetensors gguf mistral text-generation pytorch
transformers pytorch safetensors llama text-generation facebook
transformers pytorch safetensors llama text-generation facebook
transformers safetensors gemma4_assistant text-generation image-text-to-text conversational
transformers safetensors gemma4_unified_assistant text-generation any-to-any arxiv:2607.02770
transformers safetensors jais2 text-generation conversational ar
transformers gguf text-generation-inference unsloth llama trl
transformers safetensors mixtral text-generation function calling function-calling
transformers safetensors llama text-generation conversational tr
gradio region:us
transformers safetensors phi4mm text-generation nlp code
transformers pytorch llama text-generation code math
transformers safetensors mistral text-generation conversational en
This is an official implementation of the paper ``Building Math Agents with Multi-Turn Iterative Preference Learning'' with multi-turn DPO and KTO.
Proximal Policy Optimization with Beta distribution - uses multi agent Unity ML Tennis
A Continual Multi-agent RL testbed based on Hanabi
[ICML 2021] DFAC Framework: Factorizing the Value Function via Quantile Mixture for Multi-Agent Distributional Q-Learning
Source code for our paper: "ARIA: Training Language Agents with Intention-Driven Reward Aggregation".
A governed local AI build-and-memory system that trains small brains, compares them, protects the better one, archives the worse one, and preserves the evidence of why. v1.0.0/governed-v2.2.0+
transformers pytorch safetensors llama text-generation llama-2
Multi-stage LLM agent pipeline for optimizing AI kernels on Intel GPU — from analysis to autotuning.