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
jlin816/dynalang
Dynalang is a project that presents an agent capable of leveraging diverse types of language to solve tasks by predicting the future through a multimodal world model. The project is based on the paper "Learning to Model the World with Language," which was presented as an oral presentation at ICML 2024. The repository contains code implementations for training and evaluating the Dynalang agent across various environments, including HomeGrid, Messenger, Vision-and-Language Navigation (VLN), and LangRoom. Each environment requires specific setup instructions, dependencies, and datasets, which are detailed in the README. The HomeGrid environment allows training on tasks related to home settings, while Messenger involves interaction with a more complex environment requiring additional system dependencies. The VLN environment focuses on navigation tasks using the Habitat simulator and includes downloading large datasets like Matterport3D and Room-to-Room. LangRoom introduces a dictionary action space for simultaneous moving and talking, requiring a specific branch of the repository. The project also supports text pretraining and finetuning on datasets like TinyStories, enabling the agent to improve its language modeling capabilities. Training configurations are optimized for multi-GPU setups to handle the computational demands of the world model and policy training. The repository is adapted from DreamerV3, a known reinforcement learning framework, and includes scripts for running experiments, pretraining, and fine-tuning. The project aims to advance the understanding of how language can be used to model and predict the world in a multimodal context, combining language with other sensory inputs to improve task performance. It is a significant contribution to the field of language-based world modeling and embodied AI.
https://github.com/jlin816/dynalang
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.