The fastai book, published as Jupyter Notebooks

The repository provides code for running inference with the Meta Segment Anything Model 2 (SAM 2), links for downloading the trained model checkpoints, and example notebooks that show how to use th…

Reference PyTorch implementation and models for DINOv3

A unified framework for 3D content generation.

Official Code for Stable Cascade

OmniGen: Unified Image Generation. https://arxiv.org/pdf/2409.11340

OmniGen2: Exploration to Advanced Multimodal Generation.

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

A Modular Framework for 3D Gaussian Splatting and Beyond

[ICLR24] Official PyTorch Implementation of Magic123: One Image to High-Quality 3D Object Generation Using Both 2D and 3D Diffusion Priors

[CVPR 2025 Highlight] GEN3C: 3D-Informed World-Consistent Video Generation with Precise Camera Control

C3D is a modified version of BVLC caffe to support 3D ConvNets.

GaussianObject: High-Quality 3D Object Reconstruction from Four Views with Gaussian Splatting (SIGGRAPH Asia 2024, TOG)

3D mesh stylization driven by a text input in PyTorch

Roblox Foundation Model for 3D Intelligence

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

[ECCV 2024] HiDiffusion: Increases the resolution and speed of your diffusion model by only adding a single line of code!

Cosmos-Transfer1-DiffusionRenderer: High-quality video de-lighting and re-lighting based on Cosmos video diffusion framework

Repository of the paper "AnyUp: Universal Feature Upsampling".

An inference and training framework for multiple image input in Flux Kontext dev

[ECCV-2024] This is the official implementation of ZeST.

[ICLR'25] 3DTrajMaster: Mastering 3D Trajectory for Multi-Entity Motion in Video Generation

[3DV 2025 Best Paper] We present Object Images (Omages): An homage to the classic Geometry Images.

PhysX: Physical-Grounded 3D Asset Generation (NeurIPS 2025, Spotlight)

Feed-forward model for predicting 3D physics with 3DGS + NeRF

[3DV 2026] "SceneGen: Single-Image 3D Scene Generation in One Feedforward Pass"

Turn a Github Repo's contents into a big prompt for long-context models like Claude 3 Opus.

Self-reimplemented version of Long-LRM.