Generative Models

First of all, we thank following repositories for their work on high-quality image synthesis

• PGGAN
• StyleGAN
• StyleGAN2

Compared to InterFaceGAN repo, this repo optimizes the common API for generator, improves the pipeline to convert tensorflow weights to pytorch version, and involves model for StyleGAN2.

Pre-trained tensorflow weights (either officially released or trained by ourselves) can be found from following links. Please download them and save to folder pretrain/tensorflow/ before using.

NOTE: The officially released models are simply mirrored by us from the above three repositories, just in case they are not available from the official links.

PGGAN Official
Face
celebahq-1024x1024
Indoor Scene
bedroom-256x256	livingroom-256x256	diningroom-256x256	kitchen-256x256
Outdoor Scene
churchoutdoor-256x256	tower-256x256	bridge-256x256
Other Scene
restaurant-256x256	classroom-256x256	conferenceroom-256x256
Animal
person-256x256	cat-256x256	dog-256x256	bird-256x256
horse-256x256	sheep-256x256	cow-256x256
Transportation
car-256x256	bicycle-256x256	motorbike-256x256	bus-256x256
train-256x256	boat-256x256	airplane-256x256
Furniture
bottle-256x256	chair-256x256	pottedplant-256x256	tvmonitor-256x256
diningtable-256x256	sofa-256x256

StyleGAN Official
Model (Dataset)	Training Samples	Training Duration (K Images)	FID
ffhq-1024x1024	70,000	25,000	4.40
celebahq-1024x1024	30,000	25,000	5.06
bedroom-256x256	3,033,042	70,000	2.65
cat-256x256	1,657,266	70,000	8.53
car-512x384	5,520,756	46,000	3.27

StyleGAN Ours
Model (Dataset)	Training Samples	Training Duration (K Images)	FID
Face
ffhq-256x256	70,000	25,000	5.70
ffhq-512x512	70,000	25,000	5.15
LSUN Indoor Scene
livingroom-256x256	1,315,802	30,000	5.16
diningroom-256x256	657,571	25,000	4.13
kitchen-256x256	1,000,000	30,000	5.06
LSUN Indoor Scene Mixture
apartment-256x256	4 * 200,000	60,000	4.18
LSUN Outdoor Scene
churchoutdoor-256x256	126,227	30,000	4.82
tower-256x256	708,264	30,000	5.99
bridge-256x256	818,687	25,000	6.42
LSUN Other Scene
restaurant-256x256	626,331	50,000	4.03
classroom-256x256	168,103	50,000	10.10
conferenceroom-256x256	229,069	50,000	6.20

StyleGAN2 Official
Model (Dataset)	Training Samples	Training Duration (K Images)	FID
ffhq-1024x1024	70,000	25,000	2.84
church-256x256	126,227	48,000	3.86
cat-256x256	1,657,266	88,000	6.93
horse-256x256	2,000,340	100,000	3.43
car-512x384	5,520,756	57,000	2.32