Implementation of GANs in TensorFlow 2.x. Ensure you have the dependencies installed if you want to train these GANs.
$ pip install -r requirements.txt
Training GauGAN with Facades dataset
Blog post: GauGAN for conditional image generation
$ gdown https://drive.google.com/uc?id=1q4FEjQg1YSb4mPx2VdxL7LXKYu3voTMj
$ unzip -q facades_data.zip
$ python train_facades.py --facades_configs ./configs/facades.py
You can find the pre-trained checkpoints here and the training logs on Weights and Biases. You are also welcome to checkout this inference notebook (runnable on Colab) to play with the model.
[WIP] Training GauGAN with Coco-Stuff 10k
$ wget -q http://calvin.inf.ed.ac.uk/wp-content/uploads/data/cocostuffdataset/cocostuff-10k-v1.1.zip
$ unzip -q cocostuff-10k-v1.1.zip -d cocostuff10k_data
...
-
ML-GDE program that provided GCP credits for the execution of the experiments.