Compare GAN code.
This is the code that was used in "Are GANs Created Equal? A Large-Scale Study" paper (https://arxiv.org/abs/1711.10337) and in "The GAN Landscape: Losses, Architectures, Regularization, and Normalization" (https://arxiv.org/abs/1807.04720).
If you want to see the version used only in the first paper - please see the v1 branch of this repository.
The pre-trained models are available on TensorFlow Hub. Please see this colab for an example how to use them.
This repository also contains the values for the best hyperparameters for different combinations of models, regularizations and penalties.
You can see them in
generate_tasks_lib.py file and train using
To install, run:
python -m pip install -e . --user
After installing, make sure to run
It will download all the necessary datasets and frozen TF graphs. By default it will store them in
WARNING: by default this script only downloads and installs small datasets - it doesn't download celebaHQ or lsun bedrooms.
Lsun bedrooms dataset: If you want to install lsun-bedrooms you need to run t2t-datagen yourself (this dataset will take couple hours to download and unpack).
CelebaHQ dataset: currently it is not available in tensor2tensor. Please use the ProgressiveGAN github for instructions on how to prepare it.
compare_gan has two binaries:
generate_tasks- that creates a list of files with parameters to execute
run_one_task- that executes a given task, both training and evaluation, and stores results in the CSV file.
# Create tasks for experiment "test" in directory /tmp/results. See "src/generate_tasks_lib.py" to see other possible experiments. compare_gan_generate_tasks --workdir=/tmp/results --experiment=test # Run task 0 (training and eval) compare_gan_run_one_task --workdir=/tmp/results --task_num=0 --dataset_root=/tmp/datasets # Run task 1 (training and eval) compare_gan_run_one_task --workdir=/tmp/results --task_num=1 --dataset_root=/tmp/datasets
Results (FID and inception scores for checkpoints) will be stored in