Agent frameworks & infrastructure

Models, fine-tuning & training

2,504 agents · most starred first

Models, fine-tuning & training, page 6 of 26

STAR★ 66

STAR: Similarity-guided Teacher-Assisted Refinement for Super-Tiny Function Calling Models

ecoalign-forge★ 65

Multi-Agent DPO Data Synthesis Factory — 多智能体偏好训练数据自动合成框架 | 红队攻击 → 多persona审核 → 终审裁决 → DPO偏好对

CodeLlama-7b-Instruct-hf★ 65

transformers pytorch safetensors llama text-generation facebook

CodeLlama-70B-Instruct-GGUF★ 64

transformers gguf llama llama-2 text-generation code

A3C★ 63

Deep reinforcement learning using an asynchronous advantage actor-critic (A3C) model.

Hierarchical-Meta-Reinforcement-Learning★ 63

This repository contains the implementation for the paper - Exploration via Hierarchical Meta Reinforcement Learning.

Qwen3-Coder-30B-A3B-Monkey-Wrenches★ 63

Efforts toward giving Qwen 3 Coder 30B A3B proper agentic tool calling capabilities at or near 100% reliability.

splintr★ 63

A high-performance tokenizer (BPE, WordPiece, SentencePiece) built with Rust with Python bindings, focused on speed, safety, and resource optimization.

strands-env★ 63

A unified framework for building agentic RL environments.

Gemma-2-27B-Chinese-Chat★ 63

transformers safetensors gguf gemma2 text-generation llama-factory

Qwen3.6-27B-AEON-Ultimate-Uncensored-Multimodal-NVFP4-MTP-XS★ 63

transformers safetensors qwen3_5 image-text-to-text abliterated uncensored

deepseek-coder-7b-base-v1.5★ 63

transformers safetensors llama text-generation license:other text-generation-inference

verbaflow★ 62

Neural Language Model for Go

codegen-350M-multi★ 62

transformers pytorch codegen text-generation arxiv:2203.13474 license:bsd-3-clause

h2o-danube2-1.8b-chat★ 62

transformers safetensors mistral text-generation gpt llm

Qwen3-VL-MoeLORA★ 61

在千问最新的多模态image-text模型Qwen3-VL-4B-Instruct 进行多种lora微调对比效果,通过langchain+RAG+多智能体(Multi-Agent)进行部署

MARSHAL★ 60

[ICLR'26] MARSHAL: Incentivizing Multi-Agent Reasoning via Self-Play with Strategic LLMs

T2PO★ 60

【ICML2026 Spotlight】 T2PO: Uncertainty-Guided Exploration Control for Stable Multi-Turn Agentic Reinforcement Learning

CodeLlama-13B-GGUF★ 60

transformers gguf llama llama-2 text-generation code

Yi-34B-Chat-4bits★ 60

transformers pytorch safetensors llama text-generation conversational

Symbol-LLM★ 59

[ACL 2024] The project of Symbol-LLM

CodeLlama-7B-Python-GGUF★ 59

transformers gguf llama llama-2 text-generation code

North-Mini-Code-1.0-w4a16★ 59

transformers safetensors cohere2_moe text-generation conversational chat

Qwen3.6-27B-A3B-Coder-MTP-GGUF★ 59

transformers gguf code expert-pruning moe qwen3.6

RLLaVA★ 58

RLLaVA is a user-friendly framework for multi-modal RL research and optimized for resource-constrained teams.

train-tiny-llm★ 58

Train a 29M parameter GPT from Scratch

gemma-4-12B-coder-fable5-composer2.5-v1★ 58

transformers safetensors gemma4_unified image-text-to-text gemma4 coding

Qwen1.5-MoE-A2.7B-Chat-GPTQ-Int4★ 58

transformers safetensors qwen2_moe text-generation chat conversational

KakaoChatData★ 57

카카오톡 대화 데이터셋

BalanceSFT★ 57

This is the official repository of the paper "BalanceSFT: Improving LLM Function Calling with Balanced Training Signals and Data Hardness"

TinyLlama-1.1B-Chat-v0.1★ 57

transformers pytorch safetensors llama text-generation en

EMU★ 56

(Official) PyTorch implementation for Efficient Episodic Memory Utilization of Cooperative Multi-Agent Reinforcement Learning (EMU) (ICLR 2024, Oral)

baller2vecplusplus★ 56

A look-ahead multi-entity Transformer for modeling coordinated agents.

ancient-chat-llm★ 56

ancient-chat-llm: A LLM which is proficient in Chinese culture 古语说: 一个精通中国文化的大模型

MARL-LLM★ 56

LAMARL: LLM-Aided Multi-Agent Reinforcement Learning for Cooperative Policy Generation

PARL★ 56

PARL (Parallel-Agent Reinforcement Learning) is a training paradigm that teaches models to decompose complex tasks into parallel subtasks and coordinate multiple agents simultaneously.

Qwen3-4B-toolcalling-gguf-codex★ 56

safetensors gguf qwen3 function-calling tool-calling codex

ytopt★ 56

ytopt: machine-learning-based autotuning and hyperparameter optimization framework using Bayesian Optimization

h2o-danube-1.8b-chat★ 55

transformers safetensors mistral text-generation gpt llm

Dria-Agent-a-3B★ 54

transformers safetensors qwen2 text-generation code chat

LaMer★ 53

[ICLR 2026] Meta-RL Induces Exploration in Language Agents

DeepMath★ 53

Simple framework for training and evaluating math reasoning agents using local models, GRPO and vLLM.

agentlm-7b★ 53

transformers safetensors llama text-generation dataset:THUDM/AgentInstruct arxiv:2310.12823

Qwen1.5-4B-Chat★ 53

transformers safetensors qwen2 text-generation chat conversational

Agents-A1-4B-Q4_K_M-GGUF★ 53

transformers gguf quantized vlm vision agentic

DeepSeek-V4.1-Flash-1017-GSQ-agentic-strata★ 53

strata safetensors gguf deepseek_v41 deepseek deepseek-v4.1

Deep-Learning-Tensorflow★ 52

Gathers Tensorflow deep learning models.

shapley-q-learning★ 52

This repo is the implementation of paper ''SHAQ: Incorporating Shapley Value Theory into Multi-Agent Q-Learning''.

watsonx-ai-platform-demos★ 52

AI Agents, LLM Fine-tuning, Developer Productivity, Governance, IBM watsonx

Agentic-RL-Training-Recipes★ 52

Training Recipes for Agentic Reinforcement Learning in LLMs: A Survey

DOP★ 51

Codes accompanying the paper "DOP: Off-Policy Multi-Agent Decomposed Policy Gradients" (ICLR 2021, https://arxiv.org/abs/2007.12322)

transformers-to-mlx★ 51

Agent Skill to help convert transformer LLMs to mlx-lm

Agents-A1-4B-Q8_0-GGUF★ 51

transformers gguf quantized vlm vision agentic

RL2Grid★ 50

RL2Grid is a standardized benchmark for reinforcement learning (RL) agents in realistic power grid environments. Built on top of Grid2Op, it models real-time operations such as topology optimization and redispatching, with full AC power flow simulation, stochastic events, expert-informed heuristics, and safety-critical constraints.

paprika★ 50

Official Code Release for "Training a Generally Curious Agent"

D2Skill-AgenticRL★ 50

Dynamic dual-granularity skill bank for agentic RL, jointly evolving policy and skills to improve long-horizon decision making in agentic tasks.

AgentSynth★ 50

[ICLR 2026] AgentSynth: Scalable Task Generation for Generalist Computer-Use Agents

PowerFM★ 50

PowerFM is an open-source repository for foundation models in the power and energy domain. It both maintains original projects and collects community-contributed open-source projects, featuring fine-tuned and domain-trained models for tasks like load forecasting, fault detection, grid simulation, and agent control.

Qwen2.5-Coder-32B-Instruct-AWQ★ 50

transformers safetensors qwen2 text-generation code codeqwen

BEACON★ 50

[ICML 2026] Milestone-Guided Policy Learning for Long-Horizon Language Agents

qwen3.8-flash-coder-85gb-bf16★ 50

transformers safetensors qwen4_exp_text text-generation moe qwen

preference-lab★ 50

Policy-driven preference data for content moderation, with source-linked evidence, resumable runs, budget controls and TRL/ShareGPT exports.

Conditional-SeqGAN-Tensorflow★ 49

Conditional Sequence Generative Adversarial Network trained with policy gradient, Implementation in Tensorflow

multi-actor-data-selection★ 49

This is the repo for the paper Multi-Agent Collaborative Data Selection for Efficient LLM Pretraining.

MARLClassification★ 49

Image classification using reinforcement learning and multi-agent system

Function-Calling-Mistral-7B★ 49

Function Calling Mistral 7B. Learn how to make functions call for open source LLMs.

claude_converter★ 49

A zero-dependency Python module for inspecting and converting coding-agent session files (.jsonl) — Claude Code, Codex, and Pi — into the messages format expected by Hugging Face Transformers.

pytorch-transformer-chatbot★ 48

PyTorch v1.2에서 생긴 Transformer API 를 이용한 간단한 Chitchat 챗봇

Overcooked-AI★ 48

Offline Multi-Agent Reinforcement Learning Implementations: Solving Overcooked Game with Data-Driven Method

cogames★ 48

CoGames is a collection of multi-agent cooperative and competitive environments designed for reinforcement learning research.

prelude★ 48

Code for the paper "Aligning LLM Agents by Learning Latent Preference from User Edits".

DreamGym★ 48

This is AI implementation (not official) of the DreamGym framework from the paper "Scaling Agent Learning via Experience Synthesis" (arXiv:2511.03773).

Llama3-FunctionCalling★ 47

Fine-tune Llama3 model to support function calling

MAHPPO★ 47

PyTorch implementation of the paper: Multi-Agent Collaborative Inference via DNN Decoupling: Intermediate Feature Compression and Edge Learning.

SingLoRA★ 47

This repository provides a minimal, single-file implementation of SingLoRA (Single Matrix Low-Rank Adaptation) as described in the paper "SingLoRA: Low Rank Adaptation Using a Single Matrix" by Bensaïd et al.

PhoGPT-4B-Chat★ 47

transformers pytorch mpt text-generation conversational custom_code

StraTA★ 47

Implementation for the paper "StraTA: Incentivizing Agentic Reinforcement Learning with Strategic Trajectory Abstraction".

codegeex4-all-9b-GGUF★ 47

gguf glm codegeex thudm text-generation zh

T-MAAC★ 46

Transformer-based Multi-Agent Actor-Critic Framework

mgym★ 46

A collection of multi-agent reinforcement learning OpenAI gym environments

DQN_MADDPG_practice★ 46

RL projects including implementation of DQN/DDPG/MADDPG/BicNet on StarCraft II multi-agent learning environment SMAC

Quantum-Multi-Agent-Reinforcement-Learning★ 46

Quantum Multi-agent Reinforcement Learning (QMARL)

Summary-RL★ 46

Train an agent to generate high quality summaries

Skill-Entropy-RL★ 46

[NeurIPS 2026] Toward Skill-Native LLMs: Skill Entropy for Benchmarking and Training Long-Horizon Reasoning

Agent-STAR★ 46

[EMNLP 2026] Official implementation for paper "Demystifying Reinforcement Learning for Long-Horizon Tool-Using Agents: A Comprehensive Recipe"

Gemma-4-31B-Humanlike-Chat-GGUF★ 46

gguf safetensors llama.cpp gemma4 conversational roleplay

pisac★ 45

Tensorflow 2 source code for the PI-SAC agent from "Predictive Information Accelerates Learning in RL" (NeurIPS 2020)

Multi_Agent_Soft_Actor_Critic★ 45

A Pytorch Implementation of Multi Agent Soft Actor Critic

MAZero★ 45

Open-source codebase for MAZero, from "Efficient Multi-agent Reinforcement Learning by Planning" at ICLR 2024.

SocialDeductionLLM★ 45

Training and inference code for "Training Language Models for Social Deduction with Multi-Agent Reinforcement Learning"

OMIGA★ 45

[NeurIPS 2023] The official implementation of "Offline Multi-Agent Reinforcement Learning with Implicit Global-to-Local Value Regularization"

TinyLlama-1.1B-Chat-v0.4★ 45

transformers pytorch safetensors llama text-generation en

Llama3.1-70B-Chinese-Chat★ 45

transformers safetensors gguf llama text-generation llama-factory

ODIS★ 44

The implementation of ICLR 2023 paper "Discovering Generalizable Multi-agent Coordination Skills from Multi-task Offline Data".

OmModel★ 44

A collection of strong multimodal models for building multimodal AGI agents

Gato-A-Generalist-Agent★ 44

Minimal code for A Generalist Agent

async-rl★ 44

Variation of "Asynchronous Methods for Deep Reinforcement Learning" with multiple processes generating experience for agent (Keras + Theano + OpenAI Gym)[1-step Q-learning, n-step Q-learning, A3C]

granite-20b-code-instruct-8k★ 44

transformers safetensors gpt_bigcode text-generation code granite

h2o-danube3-500m-chat★ 44

transformers onnx safetensors llama text-generation gpt

ktoken★ 43

Kotlin multiplatform BPE tokenizer library for OpenAI models

More in Agent frameworks & infrastructure