Stars
GPUGrants - a list of GPU grants that I can think of
[NeurIPS 2025 Spotlight] Implementation of "KLASS: KL-Guided Fast Inference in Masked Diffusion Models"
Two Heads Are Better Than One: Audio-Visual Speech Error Correction with Dual Hypotheses
Train transformer language models with reinforcement learning.
An automatic evaluator for instruction-following language models. Human-validated, high-quality, cheap, and fast.
Fully open reproduction of DeepSeek-R1
Efficient Triton Kernels for LLM Training
Robust recipes to align language models with human and AI preferences
Official Code implementation of "Flex-Judge: Text-Only Reasoning Unleashes Zero-Shot Multimodal Evaluators"
Qwen3 is the large language model series developed by Qwen team, Alibaba Cloud.
Code for using and evaluating SpanBERT.
A high-throughput and memory-efficient inference and serving engine for LLMs
Mixture-of-Recursions: Learning Dynamic Recursive Depths for Adaptive Token-Level Computation (NeurIPS 2025)
A list of free LLM inference resources accessible via API.
verl: Volcano Engine Reinforcement Learning for LLMs
Type annotations and runtime checking for shape and dtype of JAX/NumPy/PyTorch/etc. arrays. https://docs.kidger.site/jaxtyping/
Doing simple retrieval from LLM models at various context lengths to measure accuracy
Large language models (LLMs) made easy, EasyLM is a one stop solution for pre-training, finetuning, evaluating and serving LLMs in JAX/Flax.
A PyTorch native platform for training generative AI models
[ICML 2025] XAttention: Block Sparse Attention with Antidiagonal Scoring
Gemma open-weight LLM library, from Google DeepMind
Pax is a Jax-based machine learning framework for training large scale models. Pax allows for advanced and fully configurable experimentation and parallelization, and has demonstrated industry lead…
An interactive HTML pretty-printer for machine learning research in IPython notebooks.
Modular, scalable library to train ML models
Meta Lingua: a lean, efficient, and easy-to-hack codebase to research LLMs.