Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/google/compare_gan

Compare GAN code.
https://github.com/google/compare_gan

Last synced: 4 months ago
JSON representation

Compare GAN code.

Awesome Lists containing this project

README 

        
# Compare GAN



This repository offers TensorFlow implementations for many components related to

**Generative Adversarial Networks**:



*   losses (such non-saturating GAN, least-squares GAN, and WGAN),

*   penalties (such as the gradient penalty),

*   normalization techniques (such as spectral normalization, batch

    normalization, and layer normalization),

*   neural architectures (BigGAN, ResNet, DCGAN), and

*   evaluation metrics (FID score, Inception Score, precision-recall, and KID

    score).



The code is **configurable via [Gin](https://github.com/google/gin-config)** and

runs on **GPU/TPU/CPUs**. Several research papers make use of this repository,

including:



1.  [Are GANs Created Equal? A Large-Scale Study](https://arxiv.org/abs/1711.10337)

    [[Code]](https://github.com/google/compare_gan/tree/v1)

    \

    Mario Lucic*, Karol Kurach*, Marcin Michalski, Sylvain Gelly, Olivier

    Bousquet **[NeurIPS 2018]**



2.  [The GAN Landscape: Losses, Architectures, Regularization, and Normalization](https://arxiv.org/abs/1807.04720)

    [[Code]](https://github.com/google/compare_gan/tree/v2)

    [[Colab]](https://colab.research.google.com/github/google/compare_gan/blob/v2/compare_gan/src/tfhub_models.ipynb)

    \

    Karol Kurach*, Mario Lucic*, Xiaohua Zhai, Marcin Michalski, Sylvain Gelly

    **[ICML 2019]**



3.  [Assessing Generative Models via Precision and Recall](https://arxiv.org/abs/1806.00035)

    [[Code]](https://github.com/google/compare_gan/blob/560697ee213f91048c6b4231ab79fcdd9bf20381/compare_gan/src/prd_score.py)

    \

    Mehdi S. M. Sajjadi, Olivier Bachem, Mario Lucic, Olivier Bousquet, Sylvain

    Gelly **[NeurIPS 2018]**



4.  [GILBO: One Metric to Measure Them All](https://arxiv.org/abs/1802.04874)

    [[Code]](https://github.com/google/compare_gan/blob/560697ee213f91048c6b4231ab79fcdd9bf20381/compare_gan/src/gilbo.py)

    \

    Alexander A. Alemi, Ian Fischer **[NeurIPS 2018]**



5.  [A Case for Object Compositionality in Deep Generative Models of Images](https://arxiv.org/abs/1810.10340)

    [[Code]](https://github.com/google/compare_gan/tree/v2_multigan)

    \

    Sjoerd van Steenkiste, Karol Kurach, Sylvain Gelly **[2018]**



6.  [On Self Modulation for Generative Adversarial Networks](https://arxiv.org/abs/1810.01365)

    [[Code]](https://github.com/google/compare_gan) \

    Ting Chen, Mario Lucic, Neil Houlsby, Sylvain Gelly **[ICLR 2019]**



7.  [Self-Supervised GANs via Auxiliary Rotation Loss](https://arxiv.org/abs/1811.11212)

    [[Code]](https://github.com/google/compare_gan)

    [[Colab]](https://colab.research.google.com/github/google/compare_gan/blob/v3/colabs/ssgan_demo.ipynb)

    \

    Ting Chen, Xiaohua Zhai, Marvin Ritter, Mario Lucic, Neil Houlsby **[CVPR

    2019]**



8.  [High-Fidelity Image Generation With Fewer Labels](https://arxiv.org/abs/1903.02271)

    [[Code]](https://github.com/google/compare_gan)

    [[Blog Post]](https://ai.googleblog.com/2019/03/reducing-need-for-labeled-data-in.html)

    [[Colab]](https://colab.research.google.com/github/google/compare_gan/blob/v3/colabs/s3gan_demo.ipynb)

    \

    Mario Lucic*, Michael Tschannen*, Marvin Ritter*, Xiaohua Zhai, Olivier

    Bachem, Sylvain Gelly **[ICML 2019]**



## Installation



You can easily install the library and all necessary dependencies by running:

`pip install -e .` from the `compare_gan/` folder.



## Running experiments



Simply run the `main.py` passing a `--model_dir` (this is where checkpoints are

stored) and a `--gin_config` (defines which model is trained on which data set

and other training options). We provide several example configurations in the

`example_configs/` folder:



*   **dcgan_celeba64**: DCGAN architecture with non-saturating loss on CelebA

    64x64px

*   **resnet_cifar10**: ResNet architecture with non-saturating loss and

    spectral normalization on CIFAR-10

*   **resnet_lsun-bedroom128**: ResNet architecture with WGAN loss and gradient

    penalty on LSUN-bedrooms 128x128px

*   **sndcgan_celebahq128**: SN-DCGAN architecture with non-saturating loss and

    spectral normalization on CelebA-HQ 128x128px

*   **biggan_imagenet128**: BigGAN architecture with hinge loss and spectral

    normalization on ImageNet 128x128px



### Training and evaluation



To see all available options please run `python main.py --help`. Main options:



*   To **train** the model use `--schedule=train` (default). Training is resumed

    from the last saved checkpoint.

*   To **evaluate** all checkpoints use `--schedule=continuous_eval

    --eval_every_steps=0`. To evaluate only checkpoints where the step size is

    divisible by 5000, use `--schedule=continuous_eval --eval_every_steps=5000`.

    By default, 3 averaging runs are used to estimate the Inception Score and

    the FID score. Keep in mind that when running locally on a single GPU it may

    not be possible to run training and evaluation simultaneously due to memory

    constraints.

*   To **train and evaluate** the model use `--schedule=eval_after_train

    --eval_every_steps=0`.



### Training on Cloud TPUs



We recommend using the

[ctpu tool](https://github.com/tensorflow/tpu/tree/master/tools/ctpu) to create

a Cloud TPU and corresponding Compute Engine VM. We use v3-128 Cloud TPU v3 Pod

for training models on ImageNet in 128x128 resolutions. You can use smaller

slices if you reduce the batch size (`options.batch_size` in the Gin config) or

model parameters. Keep in mind that the model quality might change. Before

training make sure that the environment variable `TPU_NAME` is set. Running

evaluation on TPUs is currently not supported. Use a VM with a single GPU

instead.



### Datasets



Compare GAN uses [TensorFlow Datasets](https://www.tensorflow.org/datasets) and

it will automatically download and prepare the data. For ImageNet you will need

to download the archive yourself. For CelebAHq you need to download and prepare

the images on your own. If you are using TPUs make sure to point the training

script to your Google Storage Bucket (`--tfds_data_dir`).