[ECCV 2024] Official implementation of the paper "Semantic-SAM: Segment and Recognize Anything at Any Granularity"

Pytorch framework for doing deep learning on point clouds.

Pointcept: Perceive the world with sparse points, a codebase for point cloud perception research. Latest works: Concerto (NeurIPS'25), Sonata (CVPR'25 Highlight), PTv3 (CVPR'24 Oral)

MambaOut: Do We Really Need Mamba for Vision? (CVPR 2025)

3D高斯论文，持续更新，欢迎交流讨论。

xDiT: A Scalable Inference Engine for Diffusion Transformers (DiTs) with Massive Parallelism

Automated dense category annotation engine that serves as the initial semantic labeling for the Segment Anything dataset (SA-1B).

[ICLR 2025] From anything to mesh like human artists. Official impl. of "MeshAnything: Artist-Created Mesh Generation with Autoregressive Transformers"

Ongoing research training transformer language models at scale, including: BERT & GPT-2

An extension of Open3D to address 3D Machine Learning tasks

A Unified Framework for Surface Reconstruction

This may be the simplest implement of DDPM. You can directly run Main.py to train the UNet on CIFAR-10 dataset and see the amazing process of denoising.

Web-based 3D visualization + Python

Muon is an optimizer for hidden layers in neural networks

MoBA: Mixture of Block Attention for Long-Context LLMs

SplaTAM: Splat, Track & Map 3D Gaussians for Dense RGB-D SLAM (CVPR 2024)

Minimalistic 4D-parallelism distributed training framework for education purpose

Autoregressive Model Beats Diffusion: 🦙 Llama for Scalable Image Generation

[ICCV 2023] Make-It-3D: High-Fidelity 3D Creation from A Single Image with Diffusion Prior

[CVPR'24 Highlight & Best Demo Award] Gaussian Splatting SLAM

[CVPR 2025] Official PyTorch Implementation of MambaVision: A Hybrid Mamba-Transformer Vision Backbone

PyTorch implementation of Pointnet2/Pointnet++

Frustum PointNets for 3D Object Detection from RGB-D Data

Meta-Transformer for Unified Multimodal Learning

[NeurIPS 2025] An official implementation of Flow-GRPO: Training Flow Matching Models via Online RL

InstantSplat: Sparse-view SfM-free Gaussian Splatting in Seconds

ProlificDreamer: High-Fidelity and Diverse Text-to-3D Generation with Variational Score Distillation (NeurIPS 2023 Spotlight)

Official repository for our work on micro-budget training of large-scale diffusion models.