Images

CIFARBaseModel	Pre-trained diffusion model on CIFAR-10 32x32.
DiTBaseModel	Pre-trained 256x256 ImageNet transformer diffusion model.
StableDiffusionBaseModel	Stable Diffusion base model.
SD15BaseModel	Pre-trained 512x512 Stable Diffusion 1.5.
SD2BaseModel	Pre-trained 512x512 Stable Diffusion 2 base.
CompressionReward	Compression reward for image models.
IncompressionReward	Incompression reward for image models.
AestheticReward	Aesthetic reward based on the aesthetic predictor from LAION.