Skip to content

The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.

dpmpp 3m + dpmpp 3m sde added

FizzleDorf3cfad03a6826
diff --git a/comfy/k_diffusion/sampling.py b/comfy/k_diffusion/sampling.py
index dd234435f..27ca7cc25 100644
--- a/comfy/k_diffusion/sampling.py
+++ b/comfy/k_diffusion/sampling.py
@@ -650,4 +650,78 @@ def sample_dpmpp_sde_gpu(model, x, sigmas, extra_args=None, callback=None, disab
noise_sampler = BrownianTreeNoiseSampler(x, sigma_min, sigma_max, seed=extra_args.get("seed", None), cpu=False) if noise_sampler is None else noise_sampler
return sample_dpmpp_sde(model, x, sigmas, extra_args=extra_args, callback=callback, disable=disable, eta=eta, s_noise=s_noise, noise_sampler=noise_sampler, r=r)
+def sample_dpmpp_3m(model, x, sigmas, extra_args=None, callback=None, disable=None, eta=1., s_noise=1., noise_sampler=None):
+ """DPM-Solver++(3M) without SDE-specific parts."""
+
+ sigma_min, sigma_max = sigmas[sigmas > 0].min(), sigmas.max()
+ noise_sampler = BrownianTreeNoiseSampler(x, sigma_min, sigma_max) if noise_sampler is None else noise_sampler
+ extra_args = {} if extra_args is None else extra_args
+ s_in = x.new_ones([x.shape[0]])
+
+ for i in trange(len(sigmas) - 1, disable=disable):
+ denoised = model(x, sigmas[i] * s_in, **extra_args)
+ if callback is not None:
+ callback({'x': x, 'i': i, 'sigma': sigmas[i], 'sigma_hat': sigmas[i], 'denoised': denoised})
+
+ # Update x using the DPM-Solver++(3M) update rule
+ t, s = -sigmas[i].log(), -sigmas[i + 1].log()
+ h = s - t
+ h_eta = h * (eta + 1)
+
+ x = torch.exp(-h_eta) * x + (-h_eta).expm1().neg() * denoised
+
+ if eta:
+ x = x + noise_sampler(sigmas[i], sigmas[i + 1]) * sigmas[i + 1] * (-2 * h * eta).expm1().neg().sqrt() * s_noise
+
+ return x
+
+def sample_dpmpp_3m_sde(model, x, sigmas, extra_args=None, callback=None, disable=None, eta=1., s_noise=1., noise_sampler=None):
+ """DPM-Solver++(3M) SDE."""
+
+ sigma_min, sigma_max = sigmas[sigmas > 0].min(), sigmas.max()
+ noise_sampler = BrownianTreeNoiseSampler(x, sigma_min, sigma_max) if noise_sampler is None else noise_sampler
+ extra_args = {} if extra_args is None else extra_args
+ s_in = x.new_ones([x.shape[0]])
+
+ denoised_1, denoised_2 = None, None
+ h_1, h_2 = None, None
+
+ for i in trange(len(sigmas) - 1, disable=disable):
+ denoised = model(x, sigmas[i] * s_in, **extra_args)
+ if callback is not None:
+ callback({'x': x, 'i': i, 'sigma': sigmas[i], 'sigma_hat': sigmas[i], 'denoised': denoised})
+ if sigmas[i + 1] == 0:
+ # Denoising step
+ x = denoised
+ else:
+ t, s = -sigmas[i].log(), -sigmas[i + 1].log()
+ h = s - t
+ h_eta = h * (eta + 1)
+
+ x = torch.exp(-h_eta) * x + (-h_eta).expm1().neg() * denoised
+
+ if h_2 is not None:
+ r0 = h_1 / h
+ r1 = h_2 / h
+ d1_0 = (denoised - denoised_1) / r0
+ d1_1 = (denoised_1 - denoised_2) / r1
+ d1 = d1_0 + (d1_0 - d1_1) * r0 / (r0 + r1)
+ d2 = (d1_0 - d1_1) / (r0 + r1)
+ phi_2 = h_eta.neg().expm1() / h_eta + 1
+ phi_3 = phi_2 / h_eta - 0.5
+ x = x + phi_2 * d1 - phi_3 * d2
+ elif h_1 is not None:
+ r = h_1 / h
+ d = (denoised - denoised_1) / r
+ phi_2 = h_eta.neg().expm1() / h_eta + 1
+ x = x + phi_2 * d
+
+ if eta:
+ x = x + noise_sampler(sigmas[i], sigmas[i + 1]) * sigmas[i + 1] * (-2 * h * eta).expm1().neg().sqrt() * s_noise
+
+ denoised_1, denoised_2 = denoised, denoised_1
+ h_1, h_2 = h, h_1
+ return x
diff --git a/comfy/samplers.py b/comfy/samplers.py
index de4f36da2..dc7c3a272 100644
--- a/comfy/samplers.py
+++ b/comfy/samplers.py
@@ -528,7 +528,7 @@ class KSampler:
SCHEDULERS = ["normal", "karras", "exponential", "simple", "ddim_uniform"]
SAMPLERS = ["euler", "euler_ancestral", "heun", "dpm_2", "dpm_2_ancestral",
"lms", "dpm_fast", "dpm_adaptive", "dpmpp_2s_ancestral", "dpmpp_sde", "dpmpp_sde_gpu",
- "dpmpp_2m", "dpmpp_2m_sde", "dpmpp_2m_sde_gpu", "ddim", "uni_pc", "uni_pc_bh2"]
+ "dpmpp_2m", "dpmpp_2m_sde", "dpmpp_2m_sde_gpu", "dpmpp_3m","dpmpp_3m_sde", "ddim", "uni_pc", "uni_pc_bh2"]
def __init__(self, model, steps, device, sampler=None, scheduler=None, denoise=None, model_options={}):
self.model = model