David

pocket-tts. Python

autoscientist-reference. AutoScientist: internal architecture & API reference (compiled from observed behavior)

mister-companion. A simple companion for MiSTer-FPGA

anime. JavaScript animation engine

flash-pi-dsv4. DeepSeek-V4-Flash on a Raspberry Pi 5 (8GB)

tilelang-puzzles. Learning TileLang with 10 puzzles!

ground-station. Browser-based ground station suite for satellite tracking, SDR reception, hardware control, and telemetry decoding

ProteinView. Terminal protein structure viewer — interactive 3D visualization of PDB/mmCIF structures with cartoon ribbons, braille rendering, and Sixel/Kitty graphics

ssd. A lightweight inference engine supporting speculative speculative decoding (SSD).

skillsbench. SkillsBench evaluates how well skills work and how effective agents are at using them.

last30days-skill. AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary

pocket-tts-onnx-export. Scripts for exporting ONNX models from the Pocket TTS PyTorch library and weights.

iris.c. Flux 2 image generation model pure C inference

ComfyUI-KJNodes. Various custom nodes for ComfyUI

claude-code-proxy. Run Claude Code on OpenAI models

erdosproblems. A community database for the problems on the erdosproblems.com site

soprano. Soprano: Instant, Ultra-Realistic Text-to-Speech

EAGLE. Official Implementation of EAGLE-1 (ICML'24), EAGLE-2 (EMNLP'24), and EAGLE-3 (NeurIPS'25).

zellij-pane-tracker. Zellij plugin that exports pane names to JSON for shell integration

ty. An extremely fast Python type checker and language server, written in Rust.

heretic. Fully automatic censorship removal for language models

Chatterbox-TTS-Extended. Modified version of Chatterbox that accepts text files as input and no character restrictions. I use it to make audiobooks, especially for my kids.

index-tts. An Industrial-Level Controllable and Efficient Zero-Shot Text-To-Speech System

chatterbox. SoTA open-source TTS

onnx-asr. A lightweight Python package for Automatic Speech Recognition using ONNX models

Z-Image. Python

rk-llama.cpp. Llama.cpp with the Rockchip NPU integration as a GGML backend.

deepseek-ocr.rs. Rust multi‑backend OCR/VLM engine (DeepSeek‑OCR-1/2, PaddleOCR‑VL, DotsOCR) with DSQ quantization and an OpenAI‑compatible server & CLI – run locally without Python.

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

gh200-Ubuntu-22.04-autoinstall. Automated installation of GH200 system using Ubuntu 22.04

Microsoft-Activation-Scripts. Open-source Windows and Office activator featuring HWID, Ohook, TSforge, and Online KMS activation methods, along with advanced troubleshooting.

serena. A powerful MCP toolkit for coding, providing semantic retrieval and editing capabilities - the IDE for your agent

Context-Engineering. "Context engineering is the delicate art and science of filling the context window with just the right information for the next step." — Andrej Karpathy. A frontier, first-principles handbook inspired by Karpathy and 3Blue1Brown for moving beyond prompt engineering to the wider discipline of context design, orchestration, and optimization.

file-agent. Python

agents-towards-production. End-to-end, code-first tutorials for building production-grade GenAI agents. From prototype to enterprise deployment.

exllamav3. An optimized quantization and inference library for running LLMs locally on modern consumer-class GPUs

2cent-tts. Dockerfile

system_prompts_leaks. Extracted system prompts from Anthropic - Claude Fable 5, Opus 5, Claude Design, Claude Code. OpenAI - ChatGPT GPT-5.6-Sol, Codex. Google - Gemini 3.5 Flash, 3.1 Pro, Antigravity. xAI - Grok, Cursor, Copilot, VS Code, Perplexity, and more. Updated regularly.

Deep-Live-Cam. real time face swap and one-click video deepfake with only a single image

12-factor-agents. What are the principles we can use to build LLM-powered software that is actually good enough to put in the hands of production customers?

awesome-RecSys. A curated list of awesome Recommender System (Books, Conferences, Researchers, Papers, Github Repositories, Useful Sites, Youtube Videos)

llama-swap. Reliable model swapping for any local OpenAI/Anthropic compatible server - llama.cpp, vllm, etc

VarNet. Python

claude-desktop-debian. Claude Desktop for Linux

fastrtc. The python library for real-time communication

claude-code-source-code-deobfuscation. This is a cleanroom deobfuscation of the official Claude Code npm package.

Orpheus-TTS. Towards Human-Sounding Speech

LatentSync. Taming Stable Diffusion for Lip Sync!

mirotalk. 🚀 Self-hosted open-source WebRTC video conferencing platform built on peer-to-peer (P2P) architecture for fast, secure real-time communication with end-to-end privacy.

glados-tts-api. OpenAI-compatible TTS API for the GlaDOS TTS model

system-prompts-and-models-of-ai-tools. FULL Augment Code, Claude Code, Cluely, CodeBuddy, Comet, Cursor, Devin AI, Junie, Kiro, Leap.new, Lovable, Manus, NotionAI, Orchids.app, Perplexity, Poke, Qoder, Replit, Same.dev, Trae, Traycer AI, VSCode Agent, Warp.dev, Windsurf, Xcode, Z.ai Code, Dia & v0. (And other Open Sourced) System Prompts, Internal Tools & AI Models

csm. A Conversational Speech Generation Model

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

DeepGEMM. DeepGEMM: clean and efficient BLAS kernel library on GPU

fast.c. Prepare for DeekSeek R1 inference: Benchmark CPU, DRAM, SSD, iGPU, GPU, ... with efficient code.

whisper-onnx-cpu. ONNX implementation of Whisper. PyTorch free.

wheatley-ai. Real-time conversational AI on ESP32-S3 using LiveKit, WebRTC and SenseCap Watcher

OnnxSlim. A Toolkit to Help Optimize Onnx Model

oci-arm-host-capacity. This script allows to bypass Oracle Cloud Infrastructure 'Out of host capacity' error immediately when additional OCI capacity will appear in your Home Region / Availability domain.

ultravox. A fast multimodal LLM for real-time voice

TabPFN. ⚡ TabPFN: Foundation Model for Tabular Data ⚡

rknputop. top-like script for rockhip NPUs on linux

YoloV8-NPU. YoloV8 NPU for the RK3566/68/88 with SAHI slicing

onnxmltools. ONNXMLTools enables conversion of models to ONNX

uv. An extremely fast Python package and project manager, written in Rust.

RKLLM-Gradio. Python

open-webui. User-friendly AI Interface (Supports Ollama, OpenAI API, ...)

wut. A terminal assistant for the hopelessly confused

onnxruntime-unity. ONNX Runtime Plugin for Unity

OpenPhonemizer. An espeak-compatible, permissively-licensed IPA phonemizer (G2P) based on DeepPhonemizer. Usable as a drop-in replacement for espeak's GPL phonemizer.

babylon. Babylon.cpp is a C and C++ library for grapheme to phoneme conversion and text to speech synthesis. For phonemization a ONNX runtime port of the DeepPhonemizer model is used. For speech synthesis VITS models are used. Piper models are compatible after a conversion script is run.

Dialogue-Datasets. A collection of plain text dialogue datasets

g2p. g2p: English Grapheme To Phoneme Conversion

MeloTTS. High-quality multi-lingual text-to-speech library by MyShell.ai. Support English, Spanish, French, Chinese, Japanese and Korean.

ChatTTS. A generative speech model for daily dialogue.

modded-nanogpt. NanoGPT (124M) in 90 seconds

CrisperWhisper. Verbatim Automatic Speech Recognition with improved word-level timestamps and filler detection

LoQT. Python

F5-TTS. Official code for "F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matching"

alignment-handbook. Robust recipes to align language models with human and AI preferences

evaluation-guidebook. Sharing both practical insights and theoretical knowledge about LLM evaluation that we gathered while managing the Open LLM Leaderboard and designing lighteval!

minGRU-pytorch. Implementation of the proposed minGRU in Pytorch

notebookLM-different-podcast.

exllamav2. A fast inference library for running LLMs locally on modern consumer-class GPUs

lm-format-enforcer. Enforce the output format (JSON Schema, Regex etc) of a language model

dspy. DSPy: The framework for programming—not prompting—language models

tianshou. An elegant PyTorch deep reinforcement learning library.

fasthtml. The fastest way to create an HTML app

aimo-progress-prize. Jupyter Notebook

reolinkapipy. Reolink Camera API written in Python 3.6

Capstan-Drive. Robotic actuator test stand utilizing a Capstan Drive reducer

GPT-SoVITS. 1 min voice data can also be used to train a good TTS model! (few shot voice cloning)

rknn-llm. Python

black. The uncompromising Python code formatter

pykan. Kolmogorov Arnold Networks

unsloth. Unsloth is a local UI for training and running Kimi K3, Gemma 4, Qwen3.6, DeepSeek, GLM and other models.

the-book-of-secret-knowledge. A collection of inspiring lists, manuals, cheatsheets, blogs, hacks, one-liners, cli/web tools and more.

build-your-own-x. Master programming by recreating your favorite technologies from scratch.

chat_templates. Chat Templates for 🤗 HuggingFace Large Language Models

ezrknn-llm. Easier usage of LLMs in Rockchip's NPU on SBCs like Orange Pi 5 and Radxa Rock 5 series

ubuntu-rockchip. Ubuntu for Rockchip RK35XX Devices

VoiceCraft. Zero-Shot Speech Editing and Text-to-Speech in the Wild

gpt-prompt-engineer. Jupyter Notebook

hf_transfer. Rust

sglang. SGLang is a high-performance serving framework for large language models and multimodal models.

GaLore. GaLore: Memory-Efficient LLM Training by Gradient Low-Rank Projection

EQ-Bench. A benchmark for emotional intelligence in large language models

safetensors. Simple, safe way to store and distribute tensors

dextro. Dataset indexer for blazing fast O(1) random access with virtually no memory overhead.

repeng. A library for making RepE control vectors

marker. Convert PDF to markdown + JSON quickly with high accuracy

meghna-hugo. Meghna Hugo is a responsive, minimalist, lightweight, and fast loading one-page business theme built with HTML5/CSS3, Hugo, JavaScript, and JQuery.

megakit-hugo-template. Business Hugo template

minbpe. Minimal, clean code for the Byte Pair Encoding (BPE) algorithm commonly used in LLM tokenization.

pillow-simd. The friendly PIL fork

PolyMind. A multimodal, function calling powered LLM webui.

metavoice-src. Foundational model for human-like, expressive TTS

lone-arena. Self-hosted LLM chatbot arena, with yourself as the only judge

microchain. function calling-based LLM agents

VALL-E-X. An open source implementation of Microsoft's VALL-E X zero-shot TTS model. Demo is available in https://plachtaa.github.io/vallex/

openinterpreter. A coding agent for open models like Kimi K3

moondream. tiny vision language model

llama.cpp. Port of Facebook's LLaMA model in C/C++

outlines. Structured Outputs

NumKong. SIMD-accelerated distances, dot products, matrix ops, geospatial & geometric kernels for 16 numeric types — from 6-bit floats to 64-bit complex — across x86, Arm, RISC-V, and WASM, with bindings for Python, Rust, C, C++, Swift, JS, and Go 📐

prometheus. [ICLR 2024 & NeurIPS 2023 WS] An Evaluator LM that is open-source, offers reproducible evaluation, and inexpensive to use. Specifically designed for fine-grained evaluation on a customized score rubric, Prometheus is a good alternative for human evaluation and GPT-4 evaluation.

tinygrad. You like pytorch? You like micrograd? You love tinygrad! ❤️

tabbyAPI. The official API server for Exllama. OAI compatible, lightweight, and fast.

lm-evaluation-harness-fast. speedup for lm-evaluation-harness; support tensor-parallel inference and data-parallel inference; support gptq, bitsandbytes, peft and exllamav2.

lm-evaluation-harness. A framework for few-shot evaluation of language models.

llama-cpp-agent. The llama-cpp-agent framework is a tool designed for easy interaction with Large Language Models (LLMs). Allowing users to chat with LLM models, execute structured function calls and get structured output. Works also with models not fine-tuned to JSON output and function calls.

logd74. A high-quality hand-curated logD7.4 dataset of 1,130 compounds

URIAL. Python

deepeval. The LLM Evaluation Framework

quip-sharp. Python

glados-tts. A GLaDOS TTS, using Forward Tacotron and HiFiGAN. Inference is fast and stable, even on the CPU. A low quality vocoder model is included for mobile use. Rudimentary TTS script included. Works perfectly on Linux, partially on Maybe someone smarter than me can make a GUI.

ComfyUI. The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface.

gpt-fast. Simple and efficient pytorch-native transformer text generation in <1000 LOC of python.

LibreChat. Enhanced ChatGPT Clone: Features Agents, MCP, Skills, DeepSeek, Anthropic, AWS, OpenAI, Responses API, Azure, Groq, o1, GPT-5, Mistral, OpenRouter, Vertex AI, Gemini, Artifacts, AI model switching, message search, Code Interpreter, langchain, DALL-E-3, OpenAPI Actions, Functions, Secure Multi-User Auth, Presets, open-source for self-hosting. Active

Video-LLaVA. 【EMNLP 2024🔥】Video-LLaVA: Learning United Visual Representation by Alignment Before Projection

cappr. Completion After Prompt Probability. Make your LLM make a choice

ctypesgen. Pure-python wrapper generator for ctypes

whisper-cpp-python. whisper.cpp bindings for python

mergekit. Tools for merging pretrained large language models.

guidance. A guidance language for controlling large language models.

llm_agents. Build agents which are controlled by LLMs

piper. A fast, local neural text to speech system

distil-whisper. Distilled variant of Whisper for speech recognition. 6x faster, 50% smaller, within 1% word error rate.

frp. A fast reverse proxy to help you expose a local server behind a NAT or firewall to the internet.

scikit-learn-intelex. Extension for Scikit-learn is a seamless way to speed up your Scikit-learn application

fine-tune-mistral. Fine-tune mistral-7B on 3090s, a100s, h100s

Stable-Diffusion-WebUI-TensorRT. TensorRT Extension for Stable Diffusion Web UI

4DGaussians. [CVPR 2024] 4D Gaussian Splatting for Real-Time Dynamic Scene Rendering

LlamaFactory. Unified Efficient Fine-Tuning of 100+ LLMs & VLMs (ACL 2024)

jetson-containers. Machine Learning Containers for NVIDIA Jetson and JetPack-L4T

ai-notes. notes for software engineers getting up to speed on new AI developments. Serves as datastore for https://latent.space writing, and product brainstorming, but has cleaned up canonical references under the /Resources folder.

FlashPerceiver. Fast and memory efficient PyTorch implementation of the Perceiver with FlashAttention.

ChatDev. ChatDev 2.0: Dev All through LLM-powered Multi-Agent Collaboration

annotated_deep_learning_paper_implementations. 🧑‍🏫 60+ Implementations/tutorials of deep learning papers with side-by-side notes 📝; including transformers (original, xl, switch, feedback, vit, ...), optimizers (adam, adabelief, sophia, ...), gans(cyclegan, stylegan2, ...), 🎮 reinforcement learning (ppo, dqn), capsnet, distillation, ... 🧠

nougat. Implementation of Nougat Neural Optical Understanding for Academic Documents

Bayesian-Flow-Networks. A simple implimentation of Bayesian Flow Networks (BFN)

llama2.py. Inference Llama 2 in one file of pure Python

USearch. Fast Open-Source Search & Clustering engine × for Vectors & Arbitrary Objects × in C++, C, Python, JavaScript, Rust, Java, Objective-C, Swift, C#, GoLang, and Wolfram 🔍

LLaMA2-Accessory. An Open-source Toolkit for LLM Development

simpleaichat. Python package for easily interfacing with chat apps, with robust features and minimal code complexity.

brc-pytorch. Implementation of the bistable recurrent cell (BRC) in PyTorch

reins. Stop AI agents from doing things you didn't ask for.

mup. maximal update parametrization (µP)

cool-retro-term. A good looking terminal emulator which mimics the old cathode display...

no-more-secrets. A command line tool that recreates the famous data decryption effect seen in the 1992 movie Sneakers.

txtai. 💡 All-in-one AI framework for semantic search, LLM orchestration and language model workflows

BetterChatGPT. An amazing UI for OpenAI's ChatGPT (Website + Windows + MacOS + Linux)

chatgpt-web. ChatGPT web interface using the OpenAI API

embetter. just a bunch of useful embeddings for scikit-learn pipelines

GPT4RoI. (ECCVW 2025)GPT4RoI: Instruction Tuning Large Language Model on Region-of-Interest

exllama. A more memory-efficient rewrite of the HF transformers implementation of Llama for use with quantized weights.

StableLM. StableLM: Stability AI Language Models

RedPajama-Data. The RedPajama-Data repository contains code for preparing large datasets for training large language models.

lit-llama. Implementation of the LLaMA language model based on nanoGPT. Supports flash attention, Int8 and GPTQ 4bit quantization, LoRA and LLaMA-Adapter fine-tuning, pre-training. Apache 2.0-licensed.

codealpaca. Python

chroma. Search infrastructure for AI

supercharger. Supercharge Open-Source AI Models

Grounded-Segment-Anything. Grounded SAM: Marrying Grounding DINO with Segment Anything & Stable Diffusion & Recognize Anything - Automatically Detect , Segment and Generate Anything

segment-anything. The repository provides code for running inference with the SegmentAnything Model (SAM), links for downloading the trained model checkpoints, and example notebooks that show how to use the model.

llama.cpp. LLM inference in C/C++

jetson_dla_tutorial. A tutorial for getting started with the Deep Learning Accelerator (DLA) on NVIDIA Jetson

serge. A web interface for chatting with Alpaca through llama.cpp. Fully dockerized, with an easy to use API.

x-transformers. A concise but complete full-attention transformer with a set of promising experimental features from various papers

faster-whisper. Faster Whisper transcription with CTranslate2

alpaca.cpp. Locally run an Instruction-Tuned Chat-Style LLM

dream-factory. Multi-threaded GUI manager for mass creation of AI-generated art with support for multiple GPUs.

alpaca-lora. Instruct-tune LLaMA on consumer hardware

llama-int8. Quantized inference code for LLaMA models

paul-graham-gpt. RAG on Paul Graham's essays.

awesome-tunneling. List of ngrok, Cloudflare Tunnel, Tailscale, and ZeroTier alternatives and other tunneling software and services. Focus on self-hosting.

RWKV-LM. RWKV (pronounced RwaKuv) is an RNN with great LLM performance, which can also be directly trained like a GPT transformer (parallelizable). We are at RWKV-7 "Goose". So it's combining the best of RNN and transformer - great performance, linear time, constant space (no kv-cache), fast training, infinite ctx_len, and free sentence embedding.

ChatRWKV. ChatRWKV is like ChatGPT but powered by RWKV (100% RNN) language model, and open source.

nanoGPT. The simplest, fastest repository for training/finetuning medium-sized GPTs.

py-webrtcvad. Python interface to the WebRTC Voice Activity Detector

awesome-implicit-representations. A curated list of resources on implicit neural representations.