Agent frameworks & infrastructure

Models, fine-tuning & training

2,504 agents · most starred first

Models, fine-tuning & training, page 22 of 26

agentgym

Agent Gym - Pytorch

agentic-learn

Self-evolving MLE agent framework

agentic-train
agentlightning

Agent Lightning — agentic RL infrastructure

agentmesh_lightning

Public Preview — Agent-Lightning RL integration for the Agent Governance Toolkit: governed training with policy enforcement

agentrl-eval

Evaluation client of AgentRL: Scaling Agentic Reinforcement Learning with a Multi-Turn, Multi-Task Framework

agentrl-worker

Task worker of AgentRL: Scaling Agentic Reinforcement Learning with a Multi-Turn, Multi-Task Framework

agents

Efficient TensorFlow implementation of reinforcement learning algorithms.

aigpt52

A Flask-based model training application with authentication.

anyGPT

A general purpose library for training any type of GPT model. Support for gpt-1, gpt-2, and gpt-3 models.

auto-learn-gpt

autoML for training and inference Deep Learning model

autoFillMaskWithCandi

Automatically mask sentences from a given input where certain words vary, and fil-mask from given candidates

autoFillMaskWithCandy

Automatically mask sentences from a given input where certain words vary, and fil-mask from given candidates

autotrain-llm

autotrain_llm

averaged-perceptron

Averaged Perceptron

aws-resource-validator-bedrock

Pydantic v2 models for AWS bedrock, shipped as a PEP 420 namespace extension of aws-resource-validator.

azure-agents-learning-sdk

Native, in-process reinforcement learning SDK for AI agents — in-memory and local-file storage by default.

babyGPT

An educational module for experimenting with unsupervised learning in large language modeling

ballml

A machine learning package with some ballsy models.

bandit-agents

Library to solve k-armed bandit problems

barrage

A supervised deep learning tool.

bedrock-models

A library that provides Bedrock models id

bellman

A Toolbox for Model-Based Reinforcement Learning in TensorFlow

bionemo-llm

BioNeMo Large Language Model Components using NeMo and Megatron

bitgpt

1 bit gpt

bitnet-dllm

BitNet 1.58-bit Masked Diffusion Language Model

bonito-llm

A lightweight library for generating synthetic instruction tuning datasets for your data without GPT.

bpe-openai

Python bindings wrapping the bpe-openai Rust tokenizer crate with a tiktoken-compatible API.

buildNanoGPT

A template for nbdev-based project

burmesegpt

BurmeseGPT is a Burmese-first AI ecosystem package, including Padauk, an agentic small language model built for tool use, function calling, and local deployment.

bytedance-tars-bytennllm

This is a reserved placeholder package to mitigate dependency confusion risks.

caft-llm
candle-llm

data represent, processing

cat-llm

The CatLLM ecosystem — LLM-powered classification for text, images, and PDFs across domains.

causal-gpt-rl

Inference runtime for Causal-GPT-RL policy bundles

causallm

Causal Large Models

cellm

Experiments with local learning instead of the global backpropagation.

charLLM

Character Level Language Models 🕺🏽

chatgpt-python

A RLFH implementation (ChatGPT) powered by ColossalAI

chatllama-py
chinese-gpt

Chinese Generative Pre-Training Transformer

claude-collector

Extract and sanitize Claude Code conversation data for training datasets

codablellm

A framework for creating and curating high-quality code datasets tailored for large language models

compressgpt-core

LLM Compression and Optimization Library - Build the smallest runnable models that preserve target accuracy

continual-foragax

A continual reinforcement learning benchmark

createllm

A Python package that enables users to create and train their own Language Learning Models (LLMs) from scratch using custom datasets. This package provides a simplified approach to building, training, and deploying custom language models tailored to specific domains or use cases.

csle-agents

Reinforcement learning agents for CSLE

cuGPT

PyTorch CUDA extension package.

curriculumagent

CurriculumAgent is a cleanup and improved version of the NeurIPS 2020 Competition Agent by Binbinchen.The agent is build to extract action sets of the Grid2Op Environment and then use rule-based agent to train a Reinforcement Learning agent.

customGPTBW

A private module for dataset loading and GPT training.

cybertron-llm

cybertron- Pytorch

dallm

Domain Adapted Language Model

ddpg-agent

Reinforcement Learning model using Deep Deterministic Policy Gradients (DDPG)

deepswarm

Neural Architecture Search Powered by Swarm Intelligence

dialogpt2

DialoGTP2 train/gen

diff-llm

LLM that predicts text diffs.

dllm-reason

DAG-guided discrete diffusion language models for reasoning

dnallm

A comprehensive toolkit for fine-tuning and inference with DNA Language Models

dragon-autodl

Implementation of the algorithmic framework from <An algorithmic framework for the optimization of deep neural networks architectures and hyperparameters>

dragon-baseline

Baseline training algorithm for the DRAGON Challenge

dragonfly-opt
dragongrad

Automatic Differentiation Library

easyagents

reinforcement learning for practitioners.

edg4llm

A unified tool to generate fine-tuning datasets for LLMs, including questions, answers, and dialogues.

ensembleswarm

Ensemble of ensembles regression utility

evolvishub-text-classification-llm

Enterprise-grade text classification library with 11+ LLM providers, streaming, monitoring, and advanced workflows

fairswarm

Provably fair particle swarm optimization for federated learning coalition selection

fastdatasets-llm

Generate high-quality LLM training datasets from documents with distillation and augmentation.

fate-llm

Federated Learning for Large Language Models

fedllm

FedLLM: A Federated Learning Library for Large Language Models.

flash-llm-rl

flash llm rl

forgellm

A comprehensive toolkit for end-to-end continued pre-training, fine-tuning, monitoring, testing and publishing of language models with MLX-LM

forte.gpt2

Provide Forte implementations of a fantastic collection of NLP tools.

fragile-gym

Gym: A universal API for reinforcement learning environments.

frame-averaging-pytorch

Frame Averaging

freesolo-agent

Thin CLI for queuing Freesolo backend training jobs.

fusellm

Train small LLMs and deploy them for fast structured extraction on CPU

g-mlp-gpt

gMLP - GPT

garage

A toolkit for reproducible reinforcement learning research

gemini-torch

Gemini - Pytorch

geminix

Gemini - Pytorch

generalized-average

Calcalating different of averages for DL pooling tasks

genesis-llm

Genesis architecture (PyTorch) and utilities for inference/benchmarking.

geoffrey-llm

A lightweight toolkit for LLM fine-tuning, inference optimization, and agent development

giotto-llm

Wrapper of many OS libraries to finetune LLMs and run them in inference mode efficiently.

glassbox-automl-agent

A transparent AutoML library built from scratch with NumPy

gpt-2-finetuning

Package for finetuning GPT-2 models

gpt-2-simple

Python package to easily retrain OpenAI's GPT-2 text-generating model on new texts.

gpt-from-scratch

Scalable GPT implementation using Distributed Data Parallel for efficient, multi-GPU training of transformer models.

gpt-neox-package

Custom extension for the transformers library including GPTNeoX models

gpt2-estimator

GPT2 in estimator

gpt2-plot

Python package to easily retrain OpenAI's GPT-2 text-generating model on new texts.

gpt2ppl-zh

Chinese sentence perplexity calculation based on GPT2 pre-trained model

gpt3-torch

GPT3 - Pytorch

gpt3_tokenizer

Encoder/Decoder and tokens counter for GPT3

gpt3datagen

Fine-Tune GPT-3 Data Generator is a python package that generate fake datasets.

gpt4-torch

GPT4 - Pytorch

gpt4o

gpt4o - Pytorch

gptbuilder

package to create datasets using the APIs of Openai and Azure(openai)

gptf

Distributed GPs using TensorFlow.

More in Agent frameworks & infrastructure