AI agents running research on single-GPU nanochat training automatically

95% token savings. 155x faster queries. 16 languages. LLMs can't read your entire codebase. TLDR extracts structure, traces dependencies, and gives them exactly what they need.

cuTile is a programming model for writing parallel kernels for NVIDIA GPUs

A framework to study AI models in Reasoning, Alignment, and use of Memory (RAM).

LLM Council works together to answer your hardest questions

Home for "How To Scale Your Model", a short blog-style textbook about scaling LLMs on TPUs

Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels

Fast CUDA matrix multiplication from scratch

🚀🤖 Crawl4AI: Open-source LLM Friendly Web Crawler & Scraper. Don't be shy, join here: https://discord.gg/jP8KfhDhyN

Prompts for our Grok chat assistant and the `@grok` bot on X.

Implementation of the sparse attention pattern proposed by the Deepseek team in their "Native Sparse Attention" paper

Fully open reproduction of DeepSeek-R1

🐙 Guides, papers, lessons, notebooks and resources for prompt engineering, context engineering, RAG, and AI Agents.

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

Code for RATIONALYST: Pre-training Process-Supervision for Improving Reasoning https://arxiv.org/pdf/2410.01044

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

A comprehensive repository of reasoning tasks for LLMs (and beyond)

LLM101n: Let's build a Storyteller

LLM Analytics

OCR, layout analysis, reading order, table recognition in 90+ languages

Convert PDF to markdown + JSON quickly with high accuracy

🔎 Monitor deep learning model training and hardware usage from your mobile phone 📱

Curate better data for LLMs

Code for Quiet-STaR

Grok open release

A multi-programming language benchmark for LLMs

MLX: An array framework for Apple silicon

DeepSeek LLM: Let there be answers