Awesome

Additions to this repo

This README will eventually be updated to correctly show how to run the following:

• Remove --outdir requirement: For train.py, generate.py, style_mixing.py, and projector.py, there is no longer a need to specify where the output will be saved. The output directories will be automatically generated, as was done in the previous implementations (ProGAN / StyleGAN / StyleGAN2).
• Setup of image snapshot resolution and its saving frequency: Use --img-snap to set how often to save the snapshot grid of fakes.jpg, as well as its resolution via --snap-res (options: 1080p, 4k, and 8k).
• Random interpolation video: Run python generate.py random-video --help on how to run it, but usage is similar as in StyleGAN2 (only the name changes).
• Style mixing videos: Run python style_mixing.py video --help on how to run it, but usage is similar as in StyleGAN2.
• Sightseeding: Interpolate between pre-selected seeds in either latent space. Run python sightseeding.py --help for an example on how to run it. Read here for more details.
• Vertical mirroring for training: Use --mirror-y=True when training your model to mirror your training images along the horizontal axis.
• Project in W+: Use --project-in-wplus when running projector.py to project in the W+ latent space. Use --help for better guidance for now. Thanks to Peter Baylies on how to do this cleanly.
  • Save all steps in the projection: Running projector.py with --save-every-step will save all the frames of the projection video as different .jpg files, as well as the projected disentangled latent vector at each step. These will be saved in the .npy format in the run dir.
• Synthesize images with your Discriminator: Run python discriminator_synthesis.py --help for generating images with the second network you have trained, the Discriminator. The following is a starting code, meant to inspire more usage of this network. Currently, there are three options available now:
  • Discriminator Dream: Inspired by DeepDream, we use our Discriminator instead of VGG. A starting image can be used, or a seed that will generate a random image. Check available options via python discriminator_synthesis.py dream --help
  • Discriminator Dream-zoom: Zoom in (or out), rotate, pan horizontally or vertically each frame; the result will be saved as a .mp4 video, which can be reversed, giving more options to the user. Check available options via python discriminator_synthesis.py dream-zoom --help
  • Visual reactive interpolation (unfinished, first step in code): Pass in a video file, and we will use the Discriminator, VGG16, or CLIP (must be installed) as an encoder of each frame into the latent space Z, which we then pass to the Generator. As it is still a work in progress, the resulting video is not that impressive with the Discriminator, as all the frames are being mapped to the same area in Z, so these fake latents should be better fine-tuned. Check available options with python discriminator_synthesis.py visual-reactive --help.
  • Style transfer: TODO Nothing here yet, will be added if the results are interesting.

Note: There's an option to --compress the generated videos using ffmpeg-python. Remember to install it via pip3 install ffmpeg-python before running any of the previous scripts, otherwise this last part of compressing the video will fail (but you will still have the video). Likewise, for the moment, we are using scikit-video for opening videos in the visual reactive interpolation. Install it via pip3 install scikit-video.

TODO:

• Update this README for a better user experience
• Create multi-crop and cut-crop for fully utilizing every image in the dataset (in lieu of center-crop).
• Automatically generate classes in a dataset for training a class-conditional StyleGAN2
• Add more interpolation options (sightseeding is done, missing circular, among others)
• Let users train rectangular models.
• Add more options for generating images using the Discriminator (Cellular Automata, use more than one layer, etc.)

StyleGAN2-ADA — Official PyTorch implementation

Training Generative Adversarial Networks with Limited Data<br> Tero Karras, Miika Aittala, Janne Hellsten, Samuli Laine, Jaakko Lehtinen, Timo Aila<br> https://arxiv.org/abs/2006.06676<br>

Abstract: Training generative adversarial networks (GAN) using too little data typically leads to discriminator overfitting, causing training to diverge. We propose an adaptive discriminator augmentation mechanism that significantly stabilizes training in limited data regimes. The approach does not require changes to loss functions or network architectures, and is applicable both when training from scratch and when fine-tuning an existing GAN on another dataset. We demonstrate, on several datasets, that good results are now possible using only a few thousand training images, often matching StyleGAN2 results with an order of magnitude fewer images. We expect this to open up new application domains for GANs. We also find that the widely used CIFAR-10 is, in fact, a limited data benchmark, and improve the record FID from 5.59 to 2.42.

For business inquiries, please contact researchinquiries@nvidia.com<br> For press and other inquiries, please contact Hector Marinez at hmarinez@nvidia.com<br>

Release notes

This repository is a faithful reimplementation of StyleGAN2-ADA in PyTorch, focusing on correctness, performance, and compatibility.

Correctness

• Full support for all primary training configurations.
• Extensive verification of image quality, training curves, and quality metrics against the TensorFlow version.
• Results are expected to match in all cases, excluding the effects of pseudo-random numbers and floating-point arithmetic.

Performance

• Training is typically 5%–30% faster compared to the TensorFlow version on NVIDIA Tesla V100 GPUs.
• Inference is up to 35% faster in high resolutions, but it may be slightly slower in low resolutions.
• GPU memory usage is comparable to the TensorFlow version.
• Faster startup time when training new networks (<50s), and also when using pre-trained networks (<4s).
• New command line options for tweaking the training performance.

Compatibility

• Compatible with old network pickles created using the TensorFlow version.
• New ZIP/PNG based dataset format for maximal interoperability with existing 3rd party tools.
• TFRecords datasets are no longer supported — they need to be converted to the new format.
• New JSON-based format for logs, metrics, and training curves.
• Training curves are also exported in the old TFEvents format if TensorBoard is installed.
• Command line syntax is mostly unchanged, with a few exceptions (e.g., dataset_tool.py).
• Comparison methods are not supported (--cmethod, --dcap, --cfg=cifarbaseline, --aug=adarv)
• Truncation is now disabled by default.

Data repository

Requirements

• Linux and Windows are supported, but we recommend Linux for performance and compatibility reasons.
• 1–8 high-end NVIDIA GPUs with at least 12 GB of memory. We have done all testing and development using NVIDIA DGX-1 with 8 Tesla V100 GPUs.
• 64-bit Python 3.7 and PyTorch 1.7.1. See https://pytorch.org/ for PyTorch install instructions.
• CUDA toolkit 11.0 or later. Use at least version 11.1 if running on RTX 3090. (Why is a separate CUDA toolkit installation required? See comments in #2.)
• Python libraries: pip install click requests tqdm pyspng ninja imageio-ffmpeg==0.4.3. We use the Anaconda3 2020.11 distribution which installs most of these by default.
• Docker users: use the provided Dockerfile to build an image with the required library dependencies.

The code relies heavily on custom PyTorch extensions that are compiled on the fly using NVCC. On Windows, the compilation requires Microsoft Visual Studio. We recommend installing Visual Studio Community Edition and adding it into PATH using "C:\Program Files (x86)\Microsoft Visual Studio\<VERSION>\Community\VC\Auxiliary\Build\vcvars64.bat".

Getting started

Pre-trained networks are stored as *.pkl files that can be referenced using local filenames or URLs:

# Generate curated MetFaces images without truncation (Fig.10 left)
python generate.py --outdir=out --trunc=1 --seeds=85,265,297,849 \
    --network=https://nvlabs-fi-cdn.nvidia.com/stylegan2-ada-pytorch/pretrained/metfaces.pkl

# Generate uncurated MetFaces images with truncation (Fig.12 upper left)
python generate.py --outdir=out --trunc=0.7 --seeds=600-605 \
    --network=https://nvlabs-fi-cdn.nvidia.com/stylegan2-ada-pytorch/pretrained/metfaces.pkl

# Generate class conditional CIFAR-10 images (Fig.17 left, Car)
python generate.py --outdir=out --seeds=0-35 --class=1 \
    --network=https://nvlabs-fi-cdn.nvidia.com/stylegan2-ada-pytorch/pretrained/cifar10.pkl

# Style mixing example
python style_mixing.py --outdir=out --rows=85,100,75,458,1500 --cols=55,821,1789,293 \
    --network=https://nvlabs-fi-cdn.nvidia.com/stylegan2-ada-pytorch/pretrained/metfaces.pkl

Outputs from the above commands are placed under out/*.png, controlled by --outdir. Downloaded network pickles are cached under $HOME/.cache/dnnlib, which can be overridden by setting the DNNLIB_CACHE_DIR environment variable. The default PyTorch extension build directory is $HOME/.cache/torch_extensions, which can be overridden by setting TORCH_EXTENSIONS_DIR.

Docker: You can run the above curated image example using Docker as follows:

docker build --tag sg2ada:latest .
./docker_run.sh python3 generate.py --outdir=out --trunc=1 --seeds=85,265,297,849 \
    --network=https://nvlabs-fi-cdn.nvidia.com/stylegan2-ada-pytorch/pretrained/metfaces.pkl

Note: The Docker image requires NVIDIA driver release r455.23 or later.

Legacy networks: The above commands can load most of the network pickles created using the previous TensorFlow versions of StyleGAN2 and StyleGAN2-ADA. However, for future compatibility, we recommend converting such legacy pickles into the new format used by the PyTorch version:

python legacy.py \
    --source=https://nvlabs-fi-cdn.nvidia.com/stylegan2/networks/stylegan2-cat-config-f.pkl \
    --dest=stylegan2-cat-config-f.pkl

Projecting images to latent space

To find the matching latent vector for a given image file, run:

python projector.py --outdir=out --target=~/mytargetimg.png \
    --network=https://nvlabs-fi-cdn.nvidia.com/stylegan2-ada-pytorch/pretrained/ffhq.pkl

For optimal results, the target image should be cropped and aligned similar to the FFHQ dataset. The above command saves the projection target out/target.png, result out/proj.png, latent vector out/projected_w.npz, and progression video out/proj.mp4. You can render the resulting latent vector by specifying --projected_w for generate.py:

python generate.py --outdir=out --projected_w=out/projected_w.npz \
    --network=https://nvlabs-fi-cdn.nvidia.com/stylegan2-ada-pytorch/pretrained/ffhq.pkl

Using networks from Python

You can use pre-trained networks in your own Python code as follows:

with open('ffhq.pkl', 'rb') as f:
    G = pickle.load(f)['G_ema'].cuda()  # torch.nn.Module
z = torch.randn([1, G.z_dim]).cuda()    # latent codes
c = None                                # class labels (not used in this example)
img = G(z, c)                           # NCHW, float32, dynamic range [-1, +1]

The above code requires torch_utils and dnnlib to be accessible via PYTHONPATH. It does not need source code for the networks themselves — their class definitions are loaded from the pickle via torch_utils.persistence.

The pickle contains three networks. 'G' and 'D' are instantaneous snapshots taken during training, and 'G_ema' represents a moving average of the generator weights over several training steps. The networks are regular instances of torch.nn.Module, with all of their parameters and buffers placed on the CPU at import and gradient computation disabled by default.

The generator consists of two submodules, G.mapping and G.synthesis, that can be executed separately. They also support various additional options:

w = G.mapping(z, c, truncation_psi=0.5, truncation_cutoff=8)
img = G.synthesis(w, noise_mode='const', force_fp32=True)

Please refer to generate.py, style_mixing.py, and projector.py for further examples.

Preparing datasets

Datasets are stored as uncompressed ZIP archives containing uncompressed PNG files and a metadata file dataset.json for labels.

Custom datasets can be created from a folder containing images; see python dataset_tool.py --help for more information. Alternatively, the folder can also be used directly as a dataset, without running it through dataset_tool.py first, but doing so may lead to suboptimal performance.

Legacy TFRecords datasets are not supported — see below for instructions on how to convert them.

FFHQ:

Step 1: Download the Flickr-Faces-HQ dataset as TFRecords.

Step 2: Extract images from TFRecords using dataset_tool.py from the TensorFlow version of StyleGAN2-ADA:

# Using dataset_tool.py from TensorFlow version at
# https://github.com/NVlabs/stylegan2-ada/
python ../stylegan2-ada/dataset_tool.py unpack \
    --tfrecord_dir=~/ffhq-dataset/tfrecords/ffhq --output_dir=/tmp/ffhq-unpacked

Step 3: Create ZIP archive using dataset_tool.py from this repository:

# Original 1024x1024 resolution.
python dataset_tool.py --source=/tmp/ffhq-unpacked --dest=~/datasets/ffhq.zip

# Scaled down 256x256 resolution.
python dataset_tool.py --source=/tmp/ffhq-unpacked --dest=~/datasets/ffhq256x256.zip \
    --width=256 --height=256

MetFaces: Download the MetFaces dataset and create ZIP archive:

python dataset_tool.py --source=~/downloads/metfaces/images --dest=~/datasets/metfaces.zip

AFHQ: Download the AFHQ dataset and create ZIP archive:

python dataset_tool.py --source=~/downloads/afhq/train/cat --dest=~/datasets/afhqcat.zip
python dataset_tool.py --source=~/downloads/afhq/train/dog --dest=~/datasets/afhqdog.zip
python dataset_tool.py --source=~/downloads/afhq/train/wild --dest=~/datasets/afhqwild.zip

CIFAR-10: Download the CIFAR-10 python version and convert to ZIP archive:

python dataset_tool.py --source=~/downloads/cifar-10-python.tar.gz --dest=~/datasets/cifar10.zip

LSUN: Download the desired categories from the LSUN project page and convert to ZIP archive:

python dataset_tool.py --source=~/downloads/lsun/raw/cat_lmdb --dest=~/datasets/lsuncat200k.zip \
    --transform=center-crop --width=256 --height=256 --max_images=200000

python dataset_tool.py --source=~/downloads/lsun/raw/car_lmdb --dest=~/datasets/lsuncar200k.zip \
    --transform=center-crop-wide --width=512 --height=384 --max_images=200000

BreCaHAD:

Step 1: Download the BreCaHAD dataset.

Step 2: Extract 512x512 resolution crops using dataset_tool.py from the TensorFlow version of StyleGAN2-ADA:

# Using dataset_tool.py from TensorFlow version at
# https://github.com/NVlabs/stylegan2-ada/
python dataset_tool.py extract_brecahad_crops --cropsize=512 \
    --output_dir=/tmp/brecahad-crops --brecahad_dir=~/downloads/brecahad/images

Step 3: Create ZIP archive using dataset_tool.py from this repository:

python dataset_tool.py --source=/tmp/brecahad-crops --dest=~/datasets/brecahad.zip

Training new networks

In its most basic form, training new networks boils down to:

python train.py --outdir=~/training-runs --data=~/mydataset.zip --gpus=1 --dry-run
python train.py --outdir=~/training-runs --data=~/mydataset.zip --gpus=1

The first command is optional; it validates the arguments, prints out the training configuration, and exits. The second command kicks off the actual training.

In this example, the results are saved to a newly created directory ~/training-runs/<ID>-mydataset-auto1, controlled by --outdir. The training exports network pickles (network-snapshot-<INT>.pkl) and example images (fakes<INT>.png) at regular intervals (controlled by --snap). For each pickle, it also evaluates FID (controlled by --metrics) and logs the resulting scores in metric-fid50k_full.jsonl (as well as TFEvents if TensorBoard is installed).

The name of the output directory reflects the training configuration. For example, 00000-mydataset-auto1 indicates that the base configuration was auto1, meaning that the hyperparameters were selected automatically for training on one GPU. The base configuration is controlled by --cfg:

The training configuration can be further customized with additional command line options:

• --aug=noaug disables ADA.
• --cond=1 enables class-conditional training (requires a dataset with labels).
• --mirror=1 amplifies the dataset with x-flips. Often beneficial, even with ADA.
• --resume=ffhq1024 --snap=10 performs transfer learning from FFHQ trained at 1024x1024.
• --resume=~/training-runs/<NAME>/network-snapshot-<INT>.pkl resumes a previous training run.
• --gamma=10 overrides R1 gamma. We recommend trying a couple of different values for each new dataset.
• --aug=ada --target=0.7 adjusts ADA target value (default: 0.6).
• --augpipe=blit enables pixel blitting but disables all other augmentations.
• --augpipe=bgcfnc enables all available augmentations (blit, geom, color, filter, noise, cutout).

Please refer to python train.py --help for the full list.

Expected training time

The total training time depends heavily on resolution, number of GPUs, dataset, desired quality, and hyperparameters. The following table lists expected wallclock times to reach different points in the training, measured in thousands of real images shown to the discriminator ("kimg"):

The above measurements were done using NVIDIA Tesla V100 GPUs with default settings (--cfg=auto --aug=ada --metrics=fid50k_full). "sec/kimg" shows the expected range of variation in raw training performance, as reported in log.txt. "GPU mem" and "CPU mem" show the highest observed memory consumption, excluding the peak at the beginning caused by torch.backends.cudnn.benchmark.

In typical cases, 25000 kimg or more is needed to reach convergence, but the results are already quite reasonable around 5000 kimg. 1000 kimg is often enough for transfer learning, which tends to converge significantly faster. The following figure shows example convergence curves for different datasets as a function of wallclock time, using the same settings as above:

Note: --cfg=auto serves as a reasonable first guess for the hyperparameters but it does not necessarily lead to optimal results for a given dataset. For example, --cfg=stylegan2 yields considerably better FID for FFHQ-140k at 1024x1024 than illustrated above. We recommend trying out at least a few different values of --gamma for each new dataset.

Quality metrics

By default, train.py automatically computes FID for each network pickle exported during training. We recommend inspecting metric-fid50k_full.jsonl (or TensorBoard) at regular intervals to monitor the training progress. When desired, the automatic computation can be disabled with --metrics=none to speed up the training slightly (3%–9%).

Additional quality metrics can also be computed after the training:

# Previous training run: look up options automatically, save result to JSONL file.
python calc_metrics.py --metrics=pr50k3_full \
    --network=~/training-runs/00000-ffhq10k-res64-auto1/network-snapshot-000000.pkl

# Pre-trained network pickle: specify dataset explicitly, print result to stdout.
python calc_metrics.py --metrics=fid50k_full --data=~/datasets/ffhq.zip --mirror=1 \
    --network=https://nvlabs-fi-cdn.nvidia.com/stylegan2-ada-pytorch/pretrained/ffhq.pkl

The first example looks up the training configuration and performs the same operation as if --metrics=pr50k3_full had been specified during training. The second example downloads a pre-trained network pickle, in which case the values of --mirror and --data must be specified explicitly.

Note that many of the metrics have a significant one-off cost when calculating them for the first time for a new dataset (up to 30min). Also note that the evaluation is done using a different random seed each time, so the results will vary if the same metric is computed multiple times.

We employ the following metrics in the ADA paper. Execution time and GPU memory usage is reported for one NVIDIA Tesla V100 GPU at 1024x1024 resolution:

In addition, the following metrics from the StyleGAN and StyleGAN2 papers are also supported:

References:

1. GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium, Heusel et al. 2017
2. Demystifying MMD GANs, Bińkowski et al. 2018
3. Improved Precision and Recall Metric for Assessing Generative Models, Kynkäänniemi et al. 2019
4. Improved Techniques for Training GANs, Salimans et al. 2016
5. A Style-Based Generator Architecture for Generative Adversarial Networks, Karras et al. 2018

License

Copyright © 2021, NVIDIA Corporation. All rights reserved.

This work is made available under the Nvidia Source Code License.

Citation

@inproceedings{Karras2020ada,
  title     = {Training Generative Adversarial Networks with Limited Data},
  author    = {Tero Karras and Miika Aittala and Janne Hellsten and Samuli Laine and Jaakko Lehtinen and Timo Aila},
  booktitle = {Proc. NeurIPS},
  year      = {2020}
}

Development

This is a research reference implementation and is treated as a one-time code drop. As such, we do not accept outside code contributions in the form of pull requests.

Acknowledgements

We thank David Luebke for helpful comments; Tero Kuosmanen and Sabu Nadarajan for their support with compute infrastructure; and Edgar Schönfeld for guidance on setting up unconditional BigGAN.