From b8921af0b9d3863fb063093bf4fbb4b1cbd7de13 Mon Sep 17 00:00:00 2001 From: Clybius Date: Wed, 3 Apr 2024 12:04:47 -0500 Subject: [PATCH] Add documentation for Supreme Sampler v1.2 update --- README.md | 19 +++++++++++++++---- 1 file changed, 15 insertions(+), 4 deletions(-) diff --git a/README.md b/README.md index 1d877ae..9896e7b 100644 --- a/README.md +++ b/README.md @@ -16,10 +16,21 @@ #### Supreme Sampler features: -* centralization: Subtracts mean from the denoised latent. This can lead to perceptually sharper results, though may change the perceivable brightness of the image. Conservatively defaults to **(0.02)**. -* normalization: Divides the denoised latent by the standard deviation. Can increase contrast in the image, though may hurt fidelity and coherency at high strengths. Conservatively defaults to **(0.01)**. -* edge_enhancement: Sharpens the latent, and then applies a bilateral blur, leaving the edges sharpened. Conservatively defaults to **(0.05)**. -* perphist: Adds previous denoised variable to the current denoised using perpendicular vector projection. Default of **(0)**, where higher values add the old denoised variable, and negative values subtract the old denoised variable. +* step_method: You have the ability to choose your own step method with this sampler! Optionally, there's a dynamic step method, which chooses the appropriate order based on the calculated error between steps, allowing you to obtain higher quality when it matters in the sampling process. Defaults to **(Euler)**. +* centralization: Subtracts mean from the denoised latent. This can lead to perceptually sharper results, though may change the perceivable brightness of the image. Conservatively defaults to **(0.05)**. +* normalization: Divides the denoised latent by the standard deviation. Can increase contrast in the image, though may hurt fidelity and coherency at high strengths. Conservatively defaults to **(0.05)**. +* edge_enhancement: Sharpens the latent, and then applies a bilateral blur, leaving the edges sharpened. Defaults to **(0.25)**. +* perphist: Adds previous denoised variable to the current denoised using perpendicular vector projection. Default of **(0.5)**, where higher values add the old denoised variable, and negative values subtract the old denoised variable. * substeps: Amount of times to iterate over each step and average the results. Can be useful for obtaining higher quality at a given step count. Inspired by [ReNoise](https://arxiv.org/pdf/2403.14602v1.pdf). Default of **(2)**. +* noise_modulation: Modulates the noise based on the denoising step or other functionality. Current options with a default of **("none")**: ("none", "intensity") +* modulation_strength: Strength of the modulation, utilizing a weighted sum between the modulated noise and normal noise. Default of **(2.0)**. Only has an effect when != "none". + +#### Supreme Sampler step methods: +* Euler: A simple 1st-order step method. +* DPM_1/2/3S: 1st, 2nd, and 3rd-order samplers of the DPM family of solvers. +* Bogacki Shampine: Denoise using the Bogacki-Shampine method of ODE solvers. 3rd-order sampler. +* RK4/RK45/Adaptive_RK: High-order Runge-Kutta samplers, where RK4 is a 4th-order sampler, RK45 is 6th-order, and Adaptive RK will choose between 4th/3rd/2nd/1st order based on error between previous and current denoised variables. +* Reversible_Heun: Utilizes an implementation of the reversible heun step method, similar to the one found in [torchsde](https://github.com/google-research/torchsde). 2nd-order sampler. +* Trapezoidal: A [method to solve ODEs](https://en.wikipedia.org/wiki/Trapezoidal_rule_(differential_equations)) derived from the [Trapezoidal Rule](https://en.wikipedia.org/wiki/Trapezoidal_rule) for computing integrals. Utilizing custom sampling within ComfyUI is encouraged for these samplers! \ No newline at end of file