• Stars
    star
    1,247
  • Rank 37,683 (Top 0.8 %)
  • Language
    Python
  • License
    Apache License 2.0
  • Created 12 months ago
  • Updated 4 months ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

Official Code for MotionCtrl [SIGGRAPH 2024]

MotionCtrl: A Unified and Flexible Motion Controller for Video Generation

πŸ‘‰ MotionCtrl for [SVD], for [VideoCrafter]

Paper   arXiv   Project Page

πŸ€— HF Demo   πŸ€— HF Demo


motionctrl_small.mp4

πŸ”₯πŸ”₯ We release the codes, models and demos for MotionCtrl on Stable Video Diffusion (SVD).

Official implementation of MotionCtrl: A Unified and Flexible Motion Controller for Video Generation.

MotionCtrl can Independently control complex camera motion and object motion of generated videos, with only a unified model.

Results of MotionCtrl+SVD

More results are in showcase_svd and our Project Page.

Results of MotionCtrl+VideoCrafter

More results are in our Project Page.


πŸ“ Changelog

  • 20231225: Release MotionCtrl deployed on LVDM/VideoCrafter.
  • 20231225: Gradio demo available. πŸ€— HF Demo
  • 20231228: Provide local gradio demo for convenience.
  • 20240115 More camera poses used for testing are provided in dataset/camera_poses
  • 20240115 Release MotionCtrl deployed on SVD. Codes are in the brach svd and Gradio Demo is available in πŸ€— HF Demo.
  • Release MotionCtrl deployed on AnimateDiff.

βš™οΈ Environment

conda create -n motionctrl python=3.10.6
conda activate motionctrl
pip install -r requirements.txt

πŸ’« Inference

  • Run local inference script

  1. Download the weights of MotionCtrl motionctrl.pth and put it to ./checkpoints.
  2. Go into configs/inference/run.sh and set condtype as 'camera_motion', 'object_motion', or 'both'.
  • condtype=camera_motion means only control the camera motion in the generated video.
  • condtype=object_motion means only control the object motion in the generated video.
  • condtype=both means control the camera motion and object motion in the generated video simultaneously.
  1. Running scripts: sh configs/inference/run.sh
  • Run local gradio demo

    python -m app --share
    

πŸ“š Citation

If you make use of our work, please cite our paper.

@inproceedings{wang2023motionctrl,
  title={MotionCtrl: A Unified and Flexible Motion Controller for Video Generation},
  author={Wang, Zhouxia and Yuan, Ziyang and Wang, Xintao and Chen, Tianshui and Xia, Menghan and Luo, Ping and Shan, Yin},
  booktitle={arXiv preprint arXiv:2312.03641},
  year={2023}
}

πŸ€— Acknowledgment

The current version of MotionCtrl is built on VideoCrafter. We appreciate the authors for sharing their awesome codebase.

❓ Contact

For any question, feel free to email [email protected] or [email protected].

More Repositories

1

GFPGAN

GFPGAN aims at developing Practical Algorithms for Real-world Face Restoration.
Python
35,397
star
2

PhotoMaker

PhotoMaker [CVPR 2024]
Jupyter Notebook
9,198
star
3

T2I-Adapter

T2I-Adapter
Python
3,377
star
4

InstantMesh

InstantMesh: Efficient 3D Mesh Generation from a Single Image with Sparse-view Large Reconstruction Models
Python
2,928
star
5

BrushNet

[ECCV 2024] The official implementation of paper "BrushNet: A Plug-and-Play Image Inpainting Model with Decomposed Dual-Branch Diffusion"
Python
1,298
star
6

MasaCtrl

[ICCV 2023] Consistent Image Synthesis and Editing
Python
699
star
7

SEED-Story

SEED-Story: Multimodal Long Story Generation with Large Language Model
Python
657
star
8

LLaMA-Pro

[ACL 2024] Progressive LLaMA with Block Expansion.
Python
459
star
9

Mix-of-Show

NeurIPS 2023, Mix-of-Show: Decentralized Low-Rank Adaptation for Multi-Concept Customization of Diffusion Models
Python
383
star
10

Open-MAGVIT2

Open-MAGVIT2: Democratizing Autoregressive Visual Generation
Python
376
star
11

AnimeSR

Codes for "AnimeSR: Learning Real-World Super-Resolution Models for Animation Videos"
Python
325
star
12

VQFR

ECCV 2022, Oral, VQFR: Blind Face Restoration with Vector-Quantized Dictionary and Parallel Decoder
Python
320
star
13

CustomNet

Python
258
star
14

SmartEdit

Official code of SmartEdit [CVPR-2024 Highlight]
Python
214
star
15

UMT

UMT is a unified and flexible framework which can handle different input modality combinations, and output video moment retrieval and/or highlight detection results.
Python
186
star
16

MM-RealSR

Codes for "Metric Learning based Interactive Modulation for Real-World Super-Resolution"
Python
152
star
17

ViT-Lens

[CVPR 2024] ViT-Lens: Towards Omni-modal Representations
Python
148
star
18

MCQ

Official code for "Bridging Video-text Retrieval with Multiple Choice Questions", CVPR 2022 (Oral).
Python
136
star
19

DeSRA

Official codes for DeSRA (ICML 2023)
Python
123
star
20

FAIG

NeurIPS 2021, Spotlight, Finding Discriminative Filters for Specific Degradations in Blind Super-Resolution
Python
118
star
21

ArcNerf

Nerf and extensions in all
Jupyter Notebook
106
star
22

ST-LLM

[ECCV 2024πŸ”₯] Official implementation of the paper "ST-LLM: Large Language Models Are Effective Temporal Learners"
Python
96
star
23

SurfelNeRF

SurfelNeRF: Neural Surfel Radiance Fields for Online Photorealistic Reconstruction of Indoor Scenes
76
star
24

RepSR

Codes for "RepSR: Training Efficient VGG-style Super-Resolution Networks with Structural Re-Parameterization and Batch Normalization"
74
star
25

mllm-npu

mllm-npu: training multimodal large language models on Ascend NPUs
Python
67
star
26

HOSNeRF

HOSNeRF: Dynamic Human-Object-Scene Neural Radiance Fields from a Single Video
Python
65
star
27

FastRealVSR

Codes for "Mitigating Artifacts in Real-World Video Super-Resolution Models"
59
star
28

ConMIM

Official codes for ConMIM (ICLR 2023)
Python
57
star
29

GVT

Official code for "What Makes for Good Visual Tokenizers for Large Language Models?".
Python
54
star
30

TVTS

Turning to Video for Transcript Sorting
Jupyter Notebook
44
star
31

BEBR

Official code for "Binary embedding based retrieval at Tencent"
Python
42
star
32

ViSFT

Python
33
star
33

pi-Tuning

Official code for "pi-Tuning: Transferring Multimodal Foundation Models with Optimal Multi-task Interpolation", ICML 2023.
Python
32
star
34

FLM

Accelerating Vision-Language Pretraining with Free Language Modeling (CVPR 2023)
Python
31
star
35

Efficient-VSR-Training

Codes for "Accelerating the Training of Video Super-Resolution"
30
star
36

DTN

Official code for "Dynamic Token Normalization Improves Vision Transformer", ICLR 2022.
Python
27
star
37

OpenCompatible

OpenCompatible provides a standard compatible training benchmark, covering practical training scenarios.
Python
24
star
38

BTS

BTS: A Bi-lingual Benchmark for Text Segmentation in the Wild
23
star
39

SGAT4PASS

This is the official implementation of the paper SGAT4PASS: Spherical Geometry-Aware Transformer for PAnoramic Semantic Segmentation (IJCAI 2023)
Python
23
star
40

SFDA

Python
20
star
41

TaCA

Official code for the paper, "TaCA: Upgrading Your Visual Foundation Model with Task-agnostic Compatible Adapter".
15
star
42

Plot2Code

Python
14
star
43

common_trainer

Common template for pytorch project. Easy to extent and modify for new project.
Python
12
star
44

TransFusion

The code repo for the ACM MM paper: TransFusion: Multi-Modal Fusion for Video Tag Inference viaTranslation-based Knowledge Embedding.
9
star
45

BasicVQ-GEN

7
star
46

ArcVis

Visualization of 3d and 2d components interactively.
Jupyter Notebook
6
star
47

VTLayout

3
star