headroomlabs-ai/headroom
Headroom is a context compression layer for AI agents, reducing token usage by 60-95% across tool outputs, logs, RAG chunks, files, and conversation history, while maintaining accuracy.
Awesome AI Agents › End-to-End Optimizers
KhoomeiK/LlamaGym
LlamaGym is a project designed to simplify the process of fine-tuning large language model (LLM) agents using online reinforcement learning (RL). Unlike traditional RL agents that learn by interacting with an environment and receiving rewards, LLM-based agents typically do not learn continuously in real time. LlamaGym addresses this gap by providing a streamlined framework that integrates LLM agents with Gym-style RL environments, enabling easier experimentation and iteration with agent prompting and hyperparameters. The core of LlamaGym is an abstract Agent class that manages the complexities involved in handling LLM conversation context, episode batching, reward assignment, and Proximal Policy Optimization (PPO) setup. This abstraction allows users to focus on defining the agent's behavior through three key methods: system prompt, observation formatting, and action extraction. Users can then instantiate their agent with a base LLM model and tokenizer, and run RL loops where the agent interacts with the environment, receives rewards, and updates its policy. LlamaGym supports environments like Blackjack from the Gym library and provides example implementations to help users get started quickly. While the project prioritizes simplicity over computational efficiency, it offers a practical entry point for researchers and developers interested in applying online RL to LLMs. The repository also references related work in the field, such as Lamorel and other projects focused on grounding LLMs with RL. Overall, LlamaGym is a weekend project and a work in progress, inviting contributions from the community. It aims to make the fine-tuning of LLM agents with reinforcement learning more accessible and manageable, facilitating advancements in the development of intelligent agents that learn from interaction in real time.
https://github.com/KhoomeiK/LlamaGym
Also in End-to-End Optimizers
topoteretes/cognee
Cognee is an open-source project that provides scalable and modular memory solutions for AI agents using ECL pipelines, enabling efficient data ingestion, knowledge graph generation, and query capabilities to enhance AI performance and reduce hallucinations.
letta-ai/letta
Letta is an open-source framework for building stateful AI agents with advanced reasoning, transparent long-term memory, and context management, supporting multiple LLM backends and providing a graphical development environment.
langchain-ai/opengpts
OpenGPTs is an open-source platform that replicates and extends OpenAI's GPTs and Assistants API, offering customizable cognitive architectures, extensive tooling, and support for multiple language models and vector databases.
aiwaves-cn/agents
Agents 2.0 is an open-source framework that enables the training and self-evolution of autonomous language agents through symbolic learning inspired by neural network methodologies.
getzep/zep
Zep is an AI memory foundation that uses a temporal Knowledge Graph to enable continuous learning and personalized experiences for AI agents by efficiently managing and retrieving contextual chat and data artifacts.
algorithmicsuperintelligence/optillm
OptiLLM is an optimizing inference proxy that significantly improves the accuracy and performance of large language models on reasoning tasks without requiring any model training, by applying over 20 advanced optimization techniques during inference.
mll-lab-nu/RAGEN
A reinforcement learning framework for training multi-turn reasoning LLM agents on the StarPO algorithm, shipping ten Gym-compatible environments plus diagnostics for echo-trap and reasoning-collapse failure modes.