Thomas Wimmer

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

captum. Model interpretability and understanding for PyTorch

SenseNova-Vision. Vision as Unified Multimodal Generation

lift4d. Implementation of Lift4d, Harmonizing Single-View 3D Estimation for 4D Reconstruction In-the-Wild, from Litman et al. of Carnegie Mellon

Surflo. Official implementation of Surflo: Consistent 3D Surface Flow Model with Global State.

smt. Pretraining Recurrent Networks without Recurrence

OmniProbe. Evaluation framework for probing the dense features of visual backbones one various tasks.

sCode. A VS Code extension to manage and monitor SLURM cluster jobs – view active jobs, browse history, cancel jobs, and submit new workloads directly from your editor.

auto-popsci. Agent Skill for generating math/science popularization videos with Manim + AI narration

neural-displacement-fields. Public code release of "Mesh Processing Non-Meshes via Neural Displacement Fields" by Yuta Noma, Zhecheng Wang, Chenxi Liu, Karan Singh, and Alec Jacobson

tuna-2. Official implementation of Tuna-2: Pixel Embeddings Beat Vision Encoders for Unified Understanding and Generation

AlbumentationsX. Next-generation Albumentations: dual-licensed for open-source and commercial use

andrej-karpathy-skills. A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.

DoRA. [ICML2024 (Oral)] Official PyTorch implementation of DoRA: Weight-Decomposed Low-Rank Adaptation

DiscreteDiffusion. Tutorial on discrete diffusion models.

research-companion. Strategic research thinking agents for Claude Code — idea evaluation, project triage, and structured brainstorming. Helps you decide which papers to write, not just how to write them.

slowrun. 100M tokens. Infinite compute. Lowest val loss wins.

penrose. Create beautiful diagrams just by typing notation in plain text.

manim. A community-maintained Python framework for creating mathematical animations.

flash-kmeans. Fast and memory-efficient exact kmeans

scaling_on_scales. When do we not need larger vision models?

eb_jepa. An open source library designed to provide community examples of Joint Embedding Predictive Architectures (JEPAs). It contains code and examples for learning representations from images, video, and action-conditioned video, as well as planning using JEPA-based models.

flashoptim. Python

flow_matching. A PyTorch library for implementing flow matching algorithms, featuring continuous and discrete flow matching implementations. It includes practical examples for both text and image modalities.

parabolic-position-encoding. Parabolic Position Encoding: Vision-Centric, Principled, Extrapolatable, General

nanojepa. jepa but nanogpt style

lightly. A python library for self-supervised learning on images.

vissl. VISSL is FAIR's library of extensible, modular and scalable components for SOTA Self-Supervised Learning with images.

zuegli. Zügli - train tickets to wallet passes

nepa. PyTorch implementation of NEPA

pixio. [CVPR 2026] Pixio: a capable vision encoder dedicated to dense prediction, simply by pixel reconstruction

Multi-View-Foundation-Models. Jupyter Notebook

tyro. CLI interfaces & config objects, from types

E-RayZer. [CVPR 2026] "E-RayZer: Self-supervised 3D Reconstruction as Spatial Visual Pre-training" official implementation.

fotowebcam.eu-downloader. A python script for downloading best-of images from www.foto-webcam.eu

FaCT. Official code for FaCT: Faithful Concept Traces for Explaining Neural Network Decisions. NeurIPS 2025

cphb. Competitive Programmer's Handbook

bgu. C++

umap. Uniform Manifold Approximation and Projection

TorchDR. TorchDR - PyTorch Dimensionality Reduction

maybe. The personal finance app for everyone

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

isochrone_maps. In this repo, I will show you how to create interactive and static isochrone maps in R using openrouteservice, leaflet, and ggplot2.

ssl-data-curation. PyTorch code for hierarchical k-means -- a data curation method for self-supervised learning

perception_models. State-of-the-art Image & Video CLIP, Multimodal Large Language Models, and More!

CLIP_benchmark. CLIP-like model evaluation

DIY-SC. Implementation for DIY-SC paper.

torch-batch-svd. A 100x faster SVD for PyTorch⚡️

faiss. A library for efficient similarity search and clustering of dense vectors.

colour. Colour Science for Python

DenseMatching. Dense matching library based on PyTorch

note. notebook archive

vit-pytorch. Implementation of Vision Transformer, a simple way to achieve SOTA in vision classification with only a single transformer encoder, in Pytorch

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

RollingDepth. [CVPR 2025] RollingDepth: Video Depth without Video Models

img-comparison-slider. Image comparison slider. Compare images before and after. Supports React, Vue, Angular.

3dgs-render-blender-addon. 3DGS Render by KIRI Engine

blendify. Lightweight Python framework that provides a high-level API for creating and rendering scenes with Blender.

neuralgcm. Hybrid ML + physics model of the Earth's atmosphere

DSINE. [CVPR 2024 Oral] Rethinking Inductive Biases for Surface Normal Estimation

penzai. A JAX research toolkit for building, editing, and visualizing neural networks.

PyViz3D. PyViz3D is a web-based visualizer for 3D objects and point clouds.

ODM. A command line toolkit to generate maps, point clouds, 3D models and DEMs from drone, balloon or kite images. 📷

syn-rep-learn. Learning from synthetic data - code and models

warp-mpm. mpm solver implemented in WARP

taichi_mpm. High-performance moving least squares material point method (MLS-MPM) solver. (ACM Transactions on Graphics, SIGGRAPH 2018)

pyFM. Python implementation of Functional Maps

PUG. This is the repository for the Photorealistic Unreal Graphics (PUG) datasets for representation learning.

potpourri3d. An invigorating blend of 3D geometry tools in Python.

clip-rocket. Code release for "Improved baselines for vision-language pre-training"

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

open_clip. An open source implementation of CLIP.

ImageBind. ImageBind One Embedding Space to Bind Them All

MinkowskiEngine. Minkowski Engine is an auto-diff neural network library for high-dimensional sparse tensors

tuning_playbook. A playbook for systematically maximizing the performance of deep learning models.

chronotrains. Shortest times between train stations in Europe

deutschland. Die wichtigsten APIs Deutschlands in einem Python Paket.

every-programmer-should-know. A collection of (mostly) technical things every software developer should know about

nnUNet. Python

torchquantum. A PyTorch-based framework for Quantum Classical Simulation, Quantum Machine Learning, Quantum Neural Networks, Parameterized Quantum Circuits with support for easy deployments on real quantum computers.

webdataset. A high-performance Python-based I/O system for large (and small) deep learning problems, with strong support for PyTorch.

SpinalNet. SpinalNet: Deep Neural Network with Gradual Input

awesome-equivariant-network. Paper list for equivariant neural network