Zach Mueller

nano-vllm. Nano vLLM

terminal-control. Rust

DeepSpec. DeepSpec: a full-stack codebase for training and evaluating speculative decoding algorithms

code. Noumena's internal version of your favorite coding tui but improved and optimized for our inference stack

nanocode. Minimal Claude Code alternative. Single Python file, zero dependencies, ~250 lines.

slop-guard. Slop Scoring to Stop Slop

lambda-cli. Unofficial CLI and MCP server for Lambda cloud GPU instances

SnapViewer. PyTorch memory allocation visualizer

reachy-mini-esp32-eyes. Reachy Mini Controllable Eyes mod. Or basis for any other hardware mod.

sonic-moe. Accelerating MoE with IO and Tile-aware Optimizations

nmoe. MoE training for Me and You and maybe other people

nanochat. The best ChatGPT that $100 can buy.

smol-moe. Python

sample_efficient_gpt. Training framework with a goal to explore the frontier of sample efficiency of small language models

swift-weights. Speeding up the loading of ML models by efficiently splitting the weights across multiple disks then reading them in parallel during model initialization

diffusion-trainer. Python

nanoMPI. Simple MPI implementation for prototyping or learning

Intuitor. [ICLR 2026] Learning to Reason without External Rewards

picoDeepResearch. Python

llm-cmd-comp-qwen. Shell completion using LLM for Qwen models

Tensor-Slayer. Tensor-Slayer : Manipulate weights and tensors of LLMs to achieve performance upgrades and introduce a novel inferenceless mechanistic interpretability

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

gpu-fryer. Where GPUs get cooked 👩‍🍳🔥

d2e. All of the JSON data files and card images that the D2eCV utilizes for Descent 2nd Edition

oura-data-visualiser. Grabs sleep data from the Oura API, stores it in a local database and allows you to visualise the data with Grafana

entropix. Entropy Based Sampling and Parallel CoT Decoding

ios-6-pods-hack. workaround for AirPod disconnection

llama-3-8b-self-align. StarCoder2-Instruct: Fully Transparent and Permissive Self-Alignment for Code Generation applied to llama 3 8b

selfcodealign. [NeurIPS'24] SelfCodeAlign: Self-Alignment for Code Generation

torchtitan. A PyTorch native platform for training generative AI models

open-gpu-kernel-modules. NVIDIA Linux open GPU with P2P support

schedule_free. Schedule-Free Optimization in PyTorch

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

pytorch-benchmark. Easily benchmark PyTorch model FLOPs, latency, throughput, allocated gpu memory and energy consumption

ChatRTX. A developer reference project for creating Retrieval Augmented Generation (RAG) chatbots on Windows using TensorRT-LLM

pytest-guide. Pytest guide for unittest users

slurm-for-dummies. A dummy's guide to setting up (and using) HPC clusters on Ubuntu 22.04LTS using Slurm and Munge. Created by the Quant Club @ UIowa.

RAG-Experiments. My learnings (publicly) on RAG systems

LLM-Workshop. LLM Workshop by Sourab Mangrulkar

transformers-in-supercomputers. Transformers training in a supercomputer with the 🤗 Stack and Slurm

PiPPy. Pipeline Parallelism for PyTorch

hyper_efficient_optimizers. Development of hyper efficient optimizers that can match/exceed AdamW, while using reduced memory

swe-study-group. Code for the SWE study group

the-art-of-debugging. The Art of Debugging Open Book

mup. maximal update parametrization (µP)

accelerate. 🚀 A simple way to launch, train, and use PyTorch models on almost any device and distributed configuration, automatic mixed precision (including fp8), and easy-to-configure FSDP and DeepSpeed support

minimal-trainer-zoo. Minimal example scripts of the Hugging Face Trainer, focused on staying under 150 lines

distributed-training-and-deepspeed. Python

pipegoose. Large scale 4D parallelism pre-training for 🤗 transformers in Mixture of Experts *(still work in progress)*

IJCAI2023-OptimalShardedDataParallel. [IJCAI2023] An automated parallel training system that combines the advantages from both data and model parallelism. If you have any interests, please visit/star/fork https://github.com/Youhe-Jiang/OptimalShardedDataParallel

docker-pytorch. A Docker image for PyTorch

nbquarto. Small python library solely for quick Quarto extensions

transformers. 🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.

datasets. 🤗 The largest hub of ready-to-use datasets for AI models with fast, easy-to-use and efficient data manipulation tools

ml-engineering. Machine Learning Engineering Open Book

namex. Clean up the public namespace of your package!

recurrent-interface-network-pytorch. Implementation of Recurrent Interface Network (RIN), for highly efficient generation of images and video without cascading networks, in Pytorch

Walk-with-fastai-revisited. Source notebook code for the course, stripped of all information. Please consider puchasing the course at https://store.walkwithfastai.com

python-patterns. A collection of design patterns/idioms in Python

github-dependents-info. Collect information about dependencies between a github repo and other repositories. Results available in JSON, markdown and badge

staticrypt. Password protect a static HTML page, decrypted in-browser in JS with no dependency. No server logic needed.

galai. Model API for GALACTICA

QuartoTemplates. HTML

course22p2. course.fast.ai 2022 part 2

fastdiffusion. Notes and plans for fastdiffusion course

fastbook2e. The fastai book, 2nd edition (in progress)

Awesome-pytorch-list. A comprehensive list of pytorch related content on github,such as different models,implementations,helper libraries,tutorials etc.

richmenu. A menu component for command line interfaces based on the Rich python library

code-hike-sample. JavaScript

nbdev2-docusaurus. Docusaurus site generation for nbdev

nbdoc-docusaurus. Create testable, reproduceable documentation with Jupyter notebooks

nb_helpers. A set of tools to work with notebooks

rich. Rich is a Python library for rich text and beautiful formatting in the terminal.

pytorch-benchmark. Open-source examination of torch distributed vs torch XLA vs single GPU

ptpt. PyTorch Personal Trainer: My framework for deep learning experiments

nbprocess-sandbox. Messing around with nbprocess

fast-map. A PyTorch implementation of computing mean average precision in parallel

diffusers. 🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch.

execnb. Execute a jupyter notebook, fast, without needing jupyter

watermark. An IPython magic extension for printing date and time stamps, version numbers, and hardware information

logos. Logos for fastai for use in websites, icons, etc.

torchdistx. Torch Distributed Experimental

course22. The fast.ai course notebooks

evaluate. 🤗 Evaluate: A library for easily evaluating machine learning models and datasets.

enumify. Jupyter Notebook

cyclemoid-pytorch. Cyclemoid implementation for PyTorch

fastreinference. A toy project

AnimeColorDeOldify. Colorise Anime/Manga Sketches with DeOldify

fastai. Documentation Sprint for the fastai deep learning library

Elixir_Huggingface_hub. A port to Elixir of Huggingface_hub scripts

docker_buddy. A lightweight interface connecting Python with Docker for superpowered commands quickly

optimum-graphcore. Blazing fast training of 🤗 Transformers on Graphcore IPUs

PyTorch-LIT. Lite Inference Toolkit (LIT) for PyTorch

DLAIE. Materials for Hawley's Deep Learning & AI Ethics course

awesome-fastai. A curated list of awesome fastai projects/blog posts/tutorials/etc.

wav2vec2-live. A live speech recognition using Facebooks wav2vec 2.0 model.

promptsource. Toolkit for creating, sharing and using natural language prompts.

Android-Image-Classifier-App. Image classifier app build in Android Studio.

data. A PyTorch repo for data loading and utilities to be shared by the PyTorch domain libraries.

fastgradio. Build fast gradio demos of fastai learners

nbprocess-old. Process and export Jupyter Notebooks fast (Jupyter not required)

nbagile. Making nbdev compatible for agile frameworks and developments

optimum. 🚀 Accelerate inference and training of 🤗 Transformers, Diffusers, TIMM and Sentence Transformers with easy to use hardware optimization tools

actions-ml-cicd. A Collection of GitHub Actions That Facilitate MLOps

wandb-action. GitHub Action That Retrieves Model Runs From Weights & Biases

tinykernel. A minimal Python kernel so you can run Python in your Python

docments. Document parameters using comments

fastdownload. Easily download, verify, and extract archives

fastconvert. A library for converting fastai version 1 code to version 2

Mixmatch-pytorch-SSL. An unofficial PyTorch implementation of MixMatch - A Holistic Approach to Semi-Supervised Learning

github-copilot-test. Python

pyprobml. Python code for "Probabilistic Machine learning" book by Kevin Murphy

saint. The official PyTorch implementation of recent paper - SAINT: Improved Neural Networks for Tabular Data via Row Attention and Contrastive Pre-Training

LovaszSoftmax. Code for the Lovász-Softmax loss (CVPR 2018)

fastai3. Research repo for code that may or may not end up in fastai3

Practical-Deep-Learning-for-Coders-2.0. Notebooks for the "A walk with fastai2" Study Group and Lecture Series

fasterai. FasterAI: Prune and Distill your models with FastAI and PyTorch

fastdebug. In-place debugger for the fastai library and pytorch

Ranger21. Ranger deep learning optimizer rewrite to use newest components

huggingface_hub. The official CLI and Python client for the Hugging Face Hub.

madgrad. MADGRAD Optimization Method

hf_examples. NLP Examples using the 🤗 libraries

fastai_minima. Minimal fastai code needed for working with pytorch

learning_fastai. Jupyter Notebook

flair. A very simple framework for state-of-the-art Natural Language Processing (NLP)

streambook. Live Python Notebooks with any Editor

nbdev-index. nbdev docs lookup for a few libraries and python itself

todotxt-mode. TypeScript

DALLE-pytorch. Implementation / replication of DALL-E, OpenAI's Text to Image Transformer, in Pytorch

python2jupyter. Convert from Python script to Jupyter notebook and vice versa

fastconda. Get packages onto your conda channel faster

kora. Convenient tools for Colab

fastapi-starter-kit. FAST API Starter Kit

torch-js. Node.js binding for PyTorch.

FastAPI-Fastai2. Starter template for image recognition server with FastAPI

NVTabular. NVTabular is a feature engineering and preprocessing library for tabular data designed to quickly and easily manipulate terabyte scale datasets used to train deep learning based recommender systems.

pypi_template. A minimal template for creating a pypi package

LibreASR. :speech_balloon: An On-Premises, Streaming Speech Recognition System

PyTorch-Soft-Argmax. PyTorch implementation of Soft-Argmax in 1D/2D/3D

GiantMIDI-Piano. Python

amazon-sagemaker-endpoint-deployment-of-fastai-model-with-torchserve. Deploy FastAI Trained PyTorch Model in TorchServe and Host in Amazon SageMaker Inference Endpoint

fastai. R interface to fast.ai

triplet-attention. Official PyTorch Implementation for "Rotate to Attend: Convolutional Triplet Attention Module." [WACV 2021]

Image-Caption-Generation-using-Fastai. Implemented image caption generation method propossed in Show, Attend, and Tell paper using the Fastai framework to describe the content of images. Achieved 24 BLEU score for Beam search size of 5. Designed a Web application for model deployment using the Flask framework.

Adabelief-Optimizer. Repository for NeurIPS 2020 Spotlight "AdaBelief Optimizer: Adapting stepsizes by the belief in observed gradients"

fastai-batch-size-finder. Implementation of OpenAI paper with Simple Noise Scale on Fastai V2

fastdoc. Create publication-quality books from Jupyter notebooks

deltapy. DeltaPy - Tabular Data Augmentation (by @firmai)

MEAL-V2. MEAL V2: Boosting Vanilla ResNet-50 to 80%+ Top-1 Accuracy on ImageNet without Tricks. In NeurIPS 2020 workshop.

medical. A Medical AI Library

UPIT. A fastai/PyTorch package for unpaired image-to-image translation.

notes. Stuff I find (see issues)

fastsetup. Setup all the things

SemTorch. Python

wireguard-fast. Simplified installation of WireGuard server for Ubuntu

self_supervised. Implementation of popular SOTA self-supervised learning algorithms as Fastai Callbacks.

icevision. An Agnostic Computer Vision Framework - Pluggable to any Training Library: Fastai, Pytorch-Lightning with more to come

UNOSAT-AI-Based-Rapid-Mapping-Service. This GitHub repository contains the machine learning models described in Edoardo Nemnni, Joseph Bullock, Samir Belabbes, Lars Bromley Fully Convolutional Neural Network for Rapid Flood Segmentation in Synthetic Aperture Radar Imagery.

fastsql1. A bit of extra usability for sqlalchemy v2.

AdahessianFastai. Jupyter Notebook

ImageNette_ImageWoof_ImageWang. Jupyter Notebook

minGPT-fastai. Fastai implementation of @karpathy's miniGPT library

fastai. The fastai deep learning library

adahessian. ADAHESSIAN: An Adaptive Second Order Optimizer for Machine Learning

gradient_blending. Jupyter Notebook

docker-containers. Docker images for fastai

Thunder-Detr. (unofficial) - customized fork of DETR, optimized for intelligent obj detection on 'real world' custom datasets

pytorch-YOLOv4. PyTorch ,ONNX and TensorRT implementation of YOLOv4

fastai-docs. Documentation for fastai

pytorch-loss. label-smooth, amsoftmax, partial-fc, focal-loss, triplet-loss, lovasz-softmax. Maybe useful

fastinference_pytorch. A PyTorch only inference wrapper for fastai

fastinference_onnx. A fastai-free onnx implementation for fastai

houghnet. [ECCV-20] Official PyTorch implementation of HoughNet, a voting-based object detector.

Siren-fastai2. Unofficial implementation of 'Implicit Neural Representations with Periodic Activation Functions'

cornell-birdcall-competition-starter-pack. Cornell Birdcall Identification (a Kaggle competition) starter pack

fastai_tpu. TPU support for the fastai library

BOSS. This repository provides the code for replicating the experiments in the paper "Building One-Shot Semi-supervised (BOSS) Learning up to Fully Supervised Performance"

yolov5. Ultralytics YOLOv5 in PyTorch for object detection, instance segmentation, classification, training, and export.

math-as-code. a cheat-sheet for mathematical notation in code form

training-detr. Unofficial Colab on how to train DETR, the intelligent object detector, with your own dataset. DETR = Detection Transformer

blurr. A library that integrates huggingface transformers with the world of fastai, giving fastai devs everything they need to train, evaluate, and deploy transformer specific models.

fastinference. A collection of inference modules for fastai2

hummingbird. Hummingbird compiles trained ML models into tensor computation for faster inference.

resnet_checkpoints. Exploring gradient chekcpoints resnets with fastai V2

amalgam. Extensions for the `fastai2` library in the form of data augmentations and inference utilities. Specific to computer vision models. Inference utilities are built with a focus on single-label (softmax) classifiers.

fastai2_tabular_hybrid. Developing and integrating methods for fastai2 tabular with other datatypes

tse. Kaggle Tweet Sentiment Extraction Competition

fastai-docker-deploy. Deploy fastai models with Docker

gen-efficientnet-pytorch. Pretrained EfficientNet, EfficientNet-Lite, MixNet, MobileNetV3 / V2, MNASNet A1 and B1, FBNet, Single-Path NAS

masks4all. Make your own mask. My mask protects you. Your mask protects me.

tfrecord. Python way to Read/Write TFRecords

course-v4. Please use fastbook's /clean folder instead of this

faststyle. Fast style transfer

nbeverything.

book.fast.ai. Information for readers of the fastai book

QuantumForest. Fast Differentiable Forest lib with the advantages of both decision trees and neural networks

fastbook. The fastai book, published as Jupyter Notebooks

timeseries. Time Series package for fastai v2

Environmental_Intelligence. Data for Environmental Intelligence: A mega list of Earth System Datasets covering earth observations, climate, water, forests, biodiversity, ecology, protected areas, natural hazards, marine and the tracking of UN's Sustainable Development Goals

CenterMask. [CVPR 2020] CenterMask : Real-Time Anchor-Free Instance Segmentation