A hash table with consistent order and fast iteration; access items by key or sequence index

whiteboard / infinite canvas SDK

Make a cascading timeline from markdown-like text. Supports simple American/European date styles, ISO8601, images, links, locations, and more.

An OAI compatible exllamav2 API that's both lightweight and fast

Personal theme for Obsidian

More relighting!

Jan is an open source alternative to ChatGPT that runs 100% offline on your computer. Multiple engine support (llama.cpp, TensorRT-LLM)

LLM training in simple, raw C/CUDA

A browser based code editor

VS Code Jupyter extension

Tile primitives for speedy kernels

A very fast linker for Linux

Pre-built implicit layer architectures with O(1) backprop, GPUs, and stiff+non-stiff DE solvers, demonstrating scientific machine learning (SciML) and physics-informed machine learning methods

Differentiable Vector Graphics Rasterization

Provide with pre-build flash-attention package wheels using GitHub Actions

Structured Text Generation

A tiny but valid `init` for containers

Mooncake is the serving platform for Kimi, a leading LLM service provided by Moonshot AI.

Material for lectures on Diffusion models at IE university

Convert PDF to markdown + JSON quickly with high accuracy

A modular graph-based Retrieval-Augmented Generation (RAG) system

🌠 Manage your shell commands.

A scheduler for GPU/CPU tasks

C-Reduce, a C and C++ program reducer

[ICML 2023] SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models

FlashInfer: Kernel Library for LLM Serving

AutoAWQ implements the AWQ algorithm for 4-bit quantization with a 2x speedup during inference. Documentation:

FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens.

Transformers-compatible library for applying various compression algorithms to LLMs for optimized deployment with vLLM

Docker implemented in around 100 lines of bash