Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

add laplace scheduler [2407.03297] #4990

Merged
merged 3 commits into from
Sep 20, 2024
Merged

add laplace scheduler [2407.03297] #4990

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
72009e7
add laplace scheduler [2407.03297]
yoinked-h Sep 20, 2024
ed42dfe
should be here instead lol
yoinked-h Sep 20, 2024
c079b69
better settings
yoinked-h Sep 20, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 11 additions & 0 deletions comfy/k_diffusion/sampling.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -44,6 +44,17 @@ def get_sigmas_vp(n, beta_d=19.9, beta_min=0.1, eps_s=1e-3, device='cpu'):
return append_zero(sigmas)


def get_sigmas_laplace(n, sigma_min, sigma_max, mu=0., beta=0.5, device='cpu'):
"""Constructs the noise schedule proposed by Tiankai et al. (2024). """
epsilon = 1e-5 # avoid log(0)
x = torch.linspace(0, 1, n, device=device)
clamp = lambda x: torch.clamp(x, min=sigma_min, max=sigma_max)
lmb = mu - beta * torch.sign(0.5-x) * torch.log(1 - 2 * torch.abs(0.5-x) + epsilon)
sigmas = clamp(torch.exp(lmb))
return sigmas



def to_d(x, sigma, denoised):
"""Converts a denoiser output to a Karras ODE derivative."""
return (x - denoised) / utils.append_dims(sigma, x.ndim)
Expand Down
22 changes: 22 additions & 0 deletions comfy_extras/nodes_custom_sampler.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -90,6 +90,27 @@ def get_sigmas(self, steps, sigma_max, sigma_min, rho):
sigmas = k_diffusion_sampling.get_sigmas_polyexponential(n=steps, sigma_min=sigma_min, sigma_max=sigma_max, rho=rho)
return (sigmas, )

class LaplaceScheduler:
@classmethod
def INPUT_TYPES(s):
return {"required":
{"steps": ("INT", {"default": 20, "min": 1, "max": 10000}),
"sigma_max": ("FLOAT", {"default": 14.614642, "min": 0.0, "max": 5000.0, "step":0.01, "round": False}),
"sigma_min": ("FLOAT", {"default": 0.0291675, "min": 0.0, "max": 5000.0, "step":0.01, "round": False}),
"mu": ("FLOAT", {"default": 0.0, "min": -10.0, "max": 10.0, "step":0.1, "round": False}),
"beta": ("FLOAT", {"default": 0.5, "min": 0.0, "max": 10.0, "step":0.1, "round": False}),
}
}
RETURN_TYPES = ("SIGMAS",)
CATEGORY = "sampling/custom_sampling/schedulers"

FUNCTION = "get_sigmas"

def get_sigmas(self, steps, sigma_max, sigma_min, mu, beta):
sigmas = k_diffusion_sampling.get_sigmas_laplace(n=steps, sigma_min=sigma_min, sigma_max=sigma_max, mu=mu, beta=beta)
return (sigmas, )


class SDTurboScheduler:
@classmethod
def INPUT_TYPES(s):
Expand Down Expand Up @@ -673,6 +694,7 @@ def add_noise(self, model, noise, sigmas, latent_image):
"KarrasScheduler": KarrasScheduler,
"ExponentialScheduler": ExponentialScheduler,
"PolyexponentialScheduler": PolyexponentialScheduler,
"LaplaceScheduler": LaplaceScheduler,
"VPScheduler": VPScheduler,
"BetaSamplingScheduler": BetaSamplingScheduler,
"SDTurboScheduler": SDTurboScheduler,
Expand Down
Loading