snap-research

Code for Motion Representations for Articulated Animation paper

EfficientFormerV2 [ICCV 2023] & EfficientFormer [NeurIPs 2022]

GRID: Generative Recommendation with Semantic IDs

[CVPR 2024] Panda-70M: Captioning 70M Videos with Multiple Cross-Modality Teachers

[ICLR 2024] Github Repo for "HyperHuman: Hyper-Realistic Human Generation with Latent Structural Diffusion"

Official implementation for "Stable Flow: Vital Layers for Training-Free Image Editing" [CVPR 2025]

[ICLR 2021 Spotlight] A Good Image Generator Is What You Need for High-Resolution Video Synthesis

3D generation on ImageNet [ICLR 2023]

[CVPR 2023] Real-Time Neural Light Field on Mobile Devices

[CVPR 2022] Show Me What and Tell Me How: Video Synthesis via Multimodal Conditioning

Official Implementation for "MyVLM: Personalizing VLMs for User-Specific Queries" (ECCV 2024)

[ECCV 2022] R2L: Distilling Neural Radiance Field to Neural Light Field for Efficient Novel View Synthesis

A dataset for 3D hand reconstruction in the wild.

[CVPR 2021] Teachers Do More Than Teach: Compressing Image-to-Image Models (CAT)

AC3D: Analyzing and Improving 3D Camera Control in Video Diffusion Transformers

[CVPR 2026] 👋 Dataset and Benchmark code for EgoEdit

Official Implementation of weights2weights

DELTA: Dense Efficient Long-range 3D Tracking for Any video (ICLR 2025)

CVPR 2023 Highlight: DiscoScene

Official Implementation for "InstantRestore: Single-Step Personalized Face Restoration with Shared-Image Attention"

Source code for the paper: "AutoDecoding Latent 3D Diffusion Models"

SnapMoGen: Human Motion Generation from Expressive Texts [NeurIPS 2025]

Official pytorch implementation of "AlphaFlow: Understanding and Improving MeanFlow Models"

This respository contains the code for the NeurIPS 2024 paper SF-V: Single Forward Video Generation Model.

[ICLR 2022] Code for Graph-less Neural Networks: Teaching Old MLPs New Tricks via Distillation (GLNN)

[ICLR 2022 Oral] F8Net: Fixed-Point 8-bit Only Multiplication for Network Quantization

Official implementation of `AToM: Amortized Text-to-Mesh using 2D Diffusion`

[ICLR 2023] MLPInit: Embarrassingly Simple GNN Training Acceleration with MLP Initialization

The repository for paper Unsupervised Volumetric Animation

GiGL is an open-source library for training and inference of Graph Neural Networks at very large (billion) scale.

Benchmark dataset and code of MSRVTT-Personalization