🔥 Turn entire websites into LLM-ready markdown or structured data. Scrape, crawl and extract with a single API.

[CVPR 2025] Magma: A Foundation Model for Multimodal AI Agents

⏰ AI conference deadline countdowns

A bidirectional pipeline parallelism algorithm for computation-communication overlap in V3/R1 training.

FlashMLA: Efficient MLA decoding kernels

Comparison of the output quality of quantization methods, using Llama 3, transformers, GGUF, EXL2.

LIMO: Less is More for Reasoning

Training-free Post-training Efficient Sub-quadratic Complexity Attention. Implemented with OpenAI Triton.

Democratizing Reinforcement Learning for LLMs

A simple tutorial to add medical reasoning using GRPO

Fasten is an open-source, self-hosted, personal/family electronic medical record aggregator, designed to integrate with 100,000's of insurances/hospitals/clinics

Self-hosted Ollama + Whisper powered AI medical scribe.

OpenHealth, AI Health Assistant | Powered by Your Data

Awesome MCP Servers - A curated list of Model Context Protocol servers

The source code of our work "Prepacking: A Simple Method for Fast Prefilling and Increased Throughput in Large Language Models"

Official codebase for "Can 1B LLM Surpass 405B LLM? Rethinking Compute-Optimal Test-Time Scaling".

Pretraining code for a large-scale depth-recurrent language model

MedXpertQA: Benchmarking Expert-Level Medical Reasoning and Understanding

Meta-Learning Initialization for Multimodal Federated Tasks

Witness the aha moment of VLM with less than $3.

Google Gen AI Python SDK provides an interface for developers to integrate Google's generative models into their Python applications.

Train to 94% on CIFAR-10 in <6.3 seconds on a single A100. Or ~95.79% in ~110 seconds (or less!)

A web framework for building highly usable healthcare applications.

Contains examples of how Open Health Stack components can be used together as the foundation for FHIR based digital health solutions

A generic proxy server for applying access-control policies for a FHIR-store.

The Android FHIR SDK is a set of Kotlin libraries for building offline-capable, mobile-first healthcare applications using the HL7® FHIR® standard on Android.

Synthetic data curation for post-training and structured data extraction

This series will take you on a journey from the fundamentals of NLP and Computer Vision to the cutting edge of Vision-Language Models.

Code for studying the super weight in LLM