# ControlVideo Official pytorch implementation of "ControlVideo: Training-free Controllable Text-to-Video Generation" [![arXiv](https://img.shields.io/badge/arXiv-2305.13077-b31b1b.svg)](https://arxiv.org/abs/2305.13077) ![visitors](https://visitor-badge.laobi.icu/badge?page_id=YBYBZhang/ControlVideo) [![Replicate](https://replicate.com/cjwbw/controlvideo/badge)](https://replicate.com/cjwbw/controlvideo)

ControlVideo adapts ControlNet to the video counterpart without any finetuning, aiming to directly inherit its high-quality and consistent generation

## News * [05/28/2023] Thanks [chenxwh](https://github.com/chenxwh), add a [Replicate demo](https://replicate.com/cjwbw/controlvideo)! * [05/25/2023] Code [ControlVideo](https://github.com/YBYBZhang/ControlVideo/) released! * [05/23/2023] Paper [ControlVideo](https://arxiv.org/abs/2305.13077) released! ## Setup ### 1. Download Weights All pre-trained weights are downloaded to `checkpoints/` directory, including the pre-trained weights of [Stable Diffusion v1.5](https://huggingface.co/runwayml/stable-diffusion-v1-5), ControlNet conditioned on [canny edges](https://huggingface.co/lllyasviel/sd-controlnet-canny), [depth maps](https://huggingface.co/lllyasviel/sd-controlnet-depth), [human poses](https://huggingface.co/lllyasviel/sd-controlnet-openpose). The `flownet.pkl` is the weights of [RIFE](https://github.com/megvii-research/ECCV2022-RIFE). The final file tree likes: ```none checkpoints ├── stable-diffusion-v1-5 ├── sd-controlnet-canny ├── sd-controlnet-depth ├── sd-controlnet-openpose ├── flownet.pkl ``` ### 2. Requirements ```shell conda create -n controlvideo python=3.10 conda activate controlvideo pip install -r requirements.txt ``` `xformers` is recommended to save memory and running time. ## Inference To perform text-to-video generation, just run this command in `inference.sh`: ```bash python inference.py \ --prompt "A striking mallard floats effortlessly on the sparkling pond." \ --condition "depth" \ --video_path "data/mallard-water.mp4" \ --output_path "outputs/" \ --video_length 15 \ --smoother_steps 19 20 \ --width 512 \ --height 512 \ # --is_long_video ``` where `--video_length` is the length of synthesized video, `--condition` represents the type of structure sequence, `--smoother_steps` determines at which timesteps to perform smoothing, and `--is_long_video` denotes whether to enable efficient long-video synthesis. ## Visualizations ### ControlVideo on depth maps


"A charming flamingo gracefully wanders in the calm and serene water, its delicate neck curving into an elegant shape."	"A striking mallard floats effortlessly on the sparkling pond."	"A gigantic yellow jeep slowly turns on a wide, smooth road in the city."

"A sleek boat glides effortlessly through the shimmering river, van gogh style."	"A majestic sailing boat cruises along the vast, azure sea."	"A contented cow ambles across the dewy, verdant pasture."

### ControlVideo on canny edges


"A young man riding a sleek, black motorbike through the winding mountain roads."	"A white swan movingon the lake, cartoon style."	"A dusty old jeep was making its way down the winding forest road, creaking and groaning with each bump and turn."

"A shiny red jeep smoothly turns on a narrow, winding road in the mountains."	"A majestic camel gracefully strides across the scorching desert sands."	"A fit man is leisurely hiking through a lush and verdant forest."

### ControlVideo on human poses


"James bond moonwalk on the beach, animation style."	"Goku in a mountain range, surreal style."	"Hulk is jumping on the street, cartoon style."	"A robot dances on a road, animation style."

### Long video generation


"A steamship on the ocean, at sunset, sketch style."	"Hulk is dancing on the beach, cartoon style."

## Citation If you make use of our work, please cite our paper. ```bibtex @article{zhang2023controlvideo, title={ControlVideo: Training-free Controllable Text-to-Video Generation}, author={Zhang, Yabo and Wei, Yuxiang and Jiang, Dongsheng and Zhang, Xiaopeng and Zuo, Wangmeng and Tian, Qi}, journal={arXiv preprint arXiv:2305.13077}, year={2023} } ``` ## Acknowledgement This work repository borrows heavily from [Diffusers](https://github.com/huggingface/diffusers), [ControlNet](https://github.com/lllyasviel/ControlNet), [Tune-A-Video](https://github.com/showlab/Tune-A-Video), and [RIFE](https://github.com/megvii-research/ECCV2022-RIFE). There are also many interesting works on video generation: [Tune-A-Video](https://github.com/showlab/Tune-A-Video), [Text2Video-Zero](https://github.com/Picsart-AI-Research/Text2Video-Zero), [Follow-Your-Pose](https://github.com/mayuelala/FollowYourPose), [Control-A-Video](https://github.com/Weifeng-Chen/control-a-video), et al.