A powerful AI model.
Random seed. Leave blank to randomize the seed
Low quality input image.
Classifier-free guidance scale for prompts.
Original churn hy-param of EDM.
Original noise hy-param of EDM.
Upsampling ratio of given inputs.
Additive positive prompt for the inputs.
Minimum resolution of output images.
Negative prompt for the inputs.
Control Strength of Stage1 (negative means invalid).
Control Strength of Stage2.
Number of steps for EDM Sampling Schedule.
Use LLaVA model to get captions.
Linearly (with sigma) increase CFG from 'spt_linear_CFG' to s_cfg.
Choose a model. SUPIR-v0Q is the default training settings with paper. SUPIR-v0F is high generalization and high image quality in most cases. Training with light degradation settings. Stage1 encoder of SUPIR-v0F remains more details when facing light degradations.
Color Fixing Type..
Start point of linearly increasing CFG.
Linearly (with sigma) increase s_stage2 from 'spt_linear_s_stage2' to s_stage2.
Start point of linearly increasing s_stage2.
Click "Generate" to create an output.