Image Generation On Imagenet 128X128

FID

評価結果

このベンチマークにおける各モデルのパフォーマンス結果

モデル名	FID	Paper Title	Repository
StyleGAN-XL	1.81	StyleGAN-XL: Scaling StyleGAN to Large Diverse Datasets
BigGAN	8.7	Large Scale GAN Training for High Fidelity Natural Image Synthesis
TarFlow	5.03	Normalizing Flows are Capable Generative Models
HiT	30.83	Improved Transformer for High-Resolution GANs
ADM-G	2.97	Diffusion Models Beat GANs on Image Synthesis
PGMGAN	21.73	Partition-Guided GANs
LeCAM + DA	6.54	Regularizing Generative Adversarial Networks under Limited Data
CDM	3.52	Cascaded Diffusion Models for High Fidelity Image Generation	-
BigGAN-deep	5.7	Large Scale GAN Training for High Fidelity Natural Image Synthesis
Poly-INR	2.08	Polynomial Implicit Neural Representations For Large Diverse Datasets
BIGRoC-pl (Guided-Diffusion)	2.77	BIGRoC: Boosting Image Generation via a Robust Classifier
VDM++	1.75	-	-
SiD2	1.26	Simpler Diffusion (SiD2): 1.5 FID on ImageNet512 with pixel-space diffusion	-
SS-GAN (sBN)	43.87	Self-Supervised GANs via Auxiliary Rotation Loss
CR-BigGAN	6.66	Consistency Regularization for Generative Adversarial Networks	-
PR-BigGAN - Recall	9.92	-	-
PR-BigGAN - Precision	20.76	-	-
BigGAN-OBRS	11.65	Optimal Budgeted Rejection Sampling for Generative Models	-
RIN	2.75	Scalable Adaptive Computation for Iterative Generation
DisCo-Diff	1.73	DisCo-Diff: Enhancing Continuous Diffusion Models with Discrete Latents

0 of 23 row(s) selected.