Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
[CVPR 2022] StyleGAN-V: A Continuous Video Generator with the Price, Image Quality and Perks of StyleGAN2
| Date | Stars |
|---|---|
| 2026-07-24 | 392 |
| 2026-07-25 | 392 |
| 2026-07-28 | 392 |
| 2026-07-30 | 392 |
| 2026-08-06 | 392 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# StyleGAN-V: A Continuous Video Generator with the Price, Image Quality and Perks of StyleGAN2
### [CVPR 2022] Official pytorch implementation
[[Project website]](https://universome.github.io/stylegan-v)
[[Paper]](https://kaust-cair.s3.amazonaws.com/stylegan-v/stylegan-v-paper.pdf)
[[Casual GAN papers summary]](https://www.casualganpapers.com/text_guided_video_editing_hd_video_generation/StyleGAN-V-explained.html?query=stylegan-v)
<div style="text-align:center">
<img src="https://user-images.githubusercontent.com/3128824/161441271-09fa5cfe-a2ae-4a7f-b5ca-ad90f5e0287e.gif" alt="Content/Motion decomposition for Face Forensics 256x256"/>
</div>
<div style="text-align:center">
<img src="https://user-images.githubusercontent.com/3128824/161441278-c7c3a43d-a3cd-417b-98c5-6b889ac32935.gif" alt="Content/Motion decomposition for Sky Timelapse 256x256"/>
</div>
Code release TODO:
- [x] Installation guide
- [x] Training code
- [x] Data preprocessing scripts
- [ ] CLIP editing scripts (50% done)
- [ ] Jupyter notebook demos
- [x] [Pre-trained checkpoints](https://disk.yandex.ru/d/v7MS7zu4mmZxXw)
## Installation
To install and activate the environment, run the following command:
```
conda env create -f environment.yaml -p env
conda activate ./env
```
For clip editing, you will need to install [StyleCLIP](https://github.com/orpatashnik/StyleCLIP) and `clip`.
This repo is built on top of [INR-GAN](https://github.com/universome/inr-gan), so make sure that it runs on your system.
If you have Ampere GPUs (A6000, A100 or RTX-3090), then use `environment-ampere.yaml` instead because it is based CUDA 11 and newer pytorch versions.
## System requirements
Our codebase uses the same system requirements as StyleGAN2-ADA: see them [here](https://github.com/NVlabs/stylegan2-ada-pytorch#requirements).
We trained all the 256x256 models on 4 V100s with 32 GB each for ~2 days.
It is very similar in training time to [StyleGAN2-ADA](https://github.com/NVlabs/stylegan2-ada-pytorch) (even a bit faster).
## Training
### Dataset structure
The dataset should be either a `.zip` archive (the default setting) or a directory structured as:
```
dataset/
video1/
- frame1.jpg
- frame2.jpg
- ...
video2/
- frame1.jpg
- frame2.jpg
- ...
...
```
We use such frame-wise structure because it makes loading faster for sparse training.
By default, we assume that the data is packed into a `.zip` archive since such representation is useful to avoid additional overhead when copying data between machines on a cluster.
You can also train from a directory: for this, just remove the `.zip` suffix from the `dataset.path` property in `configs/dataset/base.yaml`.
If you want to train on a custom dataset, then create a config for it here `configs/dataset/my_dataset_config_name.yaml` (see `configs/dataset/ffs.yaml` as an example).
The `fps` parameter is needed for visualizations purposes only, videos typically have the value of 25 or 30 FPS.
### Training StyleGAN-V
To train on FaceForensics 256x256, run:
```
python src/infra/launch.py hydra.run.dir=. exp_suffix=my_experiment_name env=local dataset=ffs dataset.resolution=256 num_gpus=4
```
To train on SkyTimelapse 256x256, run:
```
python src/infra/launch.py hydra.run.dir=. exp_suffix=my_experiment_name env=local dataset=sky_timelapse dataset.resolution=256 num_gpus=4 model.generator.time_enc.min_period_len=256
```
For SkyTimelapse 256x256, we increased the period length for the motion time encoder since the motions in this dataset are much slower/smoother, than in FaceForensics.
In practice, this parameter (and its accompanying `model.generator.motion.motion_z_distance`) influences the motion quality (but not the image quality!) the most.
If you do not want `hydra` to create some log directories (typically, you don't), add the following arguments: `hydra.output_subdir=null hydra/job_logging=disabled hydra/hydra_logging=disabled`.
In case [slurm](https://slurm.schedmd.Excerpt of 12,331 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:5ffd5f831bfbce0a, topic:video-generation, readme:video generation, readme:video editing
matched fp:5ffd5f831bfbce0a, topic:pytorch