Meta Lingua: a lean, efficient, and easy-to-hack codebase to research LLMs.

Code for running RFdiffusion

Textbook on reinforcement learning from human feedback

PyTorch native finetuning library

LeanRL is a fork of CleanRL, where selected PyTorch scripts optimized for performance using compile and cudagraphs.

A language model programming library.

Deep learning for dummies. All the practical details and useful utilities that go into working with real models.

Code used for the creation of OBELICS, an open, massive and curated collection of interleaved image-text web documents, containing 141M documents, 115B text tokens and 353M images.

Dermatology ddx dataset, Jax implementations of Monte Carlo conformal prediction, plausibility regions and statistical annotation aggregation from our recent work on uncertain ground truth (TMLR'23…

A free, self-hostable news aggregator…

Trace, the New AutoDiff for AI Systems and LLM Agents

Hybrid ML + physics model of the Earth's atmosphere

RAGFlow is an open-source RAG (Retrieval-Augmented Generation) engine based on deep document understanding.

[ICML 2024] LLMCompiler: An LLM Compiler for Parallel Function Calling

Python inverse kinematics based on MuJoCo

Rapidly build AI apps in Python

LOTUS: The semantic query engine - process data with LLMs as easily as writing pandas code

A pure NumPy implementation of Mamba.

world modeling challenge for humanoid robots

Convert PDF to markdown quickly with high accuracy

Burn is a new comprehensive dynamic Deep Learning Framework built using Rust with extreme flexibility, compute efficiency and portability as its primary goals.

Scalable data pre processing and curation toolkit for LLMs

Analyzing and Improving the Training Dynamics of Diffusion Models (EDM2)