The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.
Gpu variant of dpmpp_3m_sde. Note: use 3m with exponential or karras.
comfyanonymous58c7da366572
diff --git a/comfy/k_diffusion/sampling.py b/comfy/k_diffusion/sampling.pyindex 27ca7cc25..020e65ada 100644--- a/comfy/k_diffusion/sampling.py+++ b/comfy/k_diffusion/sampling.py@@ -631,25 +631,13 @@ def sample_dpmpp_2m_sde(model, x, sigmas, extra_args=None, callback=None, disablelif solver_type == 'midpoint':x = x + 0.5 * (-h - eta_h).expm1().neg() * (1 / r) * (denoised - old_denoised)- x = x + noise_sampler(sigmas[i], sigmas[i + 1]) * sigmas[i + 1] * (-2 * eta_h).expm1().neg().sqrt() * s_noise+ if eta:+ x = x + noise_sampler(sigmas[i], sigmas[i + 1]) * sigmas[i + 1] * (-2 * eta_h).expm1().neg().sqrt() * s_noiseold_denoised = denoisedh_last = hreturn x[email protected]_grad()-def sample_dpmpp_2m_sde_gpu(model, x, sigmas, extra_args=None, callback=None, disable=None, eta=1., s_noise=1., noise_sampler=None, solver_type='midpoint'):- sigma_min, sigma_max = sigmas[sigmas > 0].min(), sigmas.max()- 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_2m_sde(model, x, sigmas, extra_args=extra_args, callback=callback, disable=disable, eta=eta, s_noise=s_noise, noise_sampler=noise_sampler, solver_type=solver_type)--[email protected]_grad()-def sample_dpmpp_sde_gpu(model, x, sigmas, extra_args=None, callback=None, disable=None, eta=1., s_noise=1., noise_sampler=None, r=1 / 2):- sigma_min, sigma_max = sigmas[sigmas > 0].min(), sigmas.max()- 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)-@torch.no_grad()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."""@@ -663,7 +651,7 @@ def sample_dpmpp_3m(model, x, sigmas, extra_args=None, callback=None, disable=Nodenoised = 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 rulet, s = -sigmas[i].log(), -sigmas[i + 1].log()h = s - t@@ -680,8 +668,9 @@ def sample_dpmpp_3m(model, x, sigmas, extra_args=None, callback=None, disable=Nodef 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."""+ seed = extra_args.get("seed", None)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+ noise_sampler = BrownianTreeNoiseSampler(x, sigma_min, sigma_max, seed=seed, cpu=True) if noise_sampler is None else noise_samplerextra_args = {} if extra_args is None else extra_argss_in = x.new_ones([x.shape[0]])@@ -725,3 +714,21 @@ def sample_dpmpp_3m_sde(model, x, sigmas, extra_args=None, callback=None, disablh_1, h_2 = h, h_1return x[email protected]_grad()+def sample_dpmpp_3m_sde_gpu(model, x, sigmas, extra_args=None, callback=None, disable=None, eta=1., s_noise=1., noise_sampler=None):+ sigma_min, sigma_max = sigmas[sigmas > 0].min(), sigmas.max()+ 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_3m_sde(model, x, sigmas, extra_args=extra_args, callback=callback, disable=disable, eta=eta, s_noise=s_noise, noise_sampler=noise_sampler)+[email protected]_grad()+def sample_dpmpp_2m_sde_gpu(model, x, sigmas, extra_args=None, callback=None, disable=None, eta=1., s_noise=1., noise_sampler=None, solver_type='midpoint'):+ sigma_min, sigma_max = sigmas[sigmas > 0].min(), sigmas.max()+ 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_2m_sde(model, x, sigmas, extra_args=extra_args, callback=callback, disable=disable, eta=eta, s_noise=s_noise, noise_sampler=noise_sampler, solver_type=solver_type)+[email protected]_grad()+def sample_dpmpp_sde_gpu(model, x, sigmas, extra_args=None, callback=None, disable=None, eta=1., s_noise=1., noise_sampler=None, r=1 / 2):+ sigma_min, sigma_max = sigmas[sigmas > 0].min(), sigmas.max()+ 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)+diff --git a/comfy/samplers.py b/comfy/samplers.pyindex dc7c3a272..1bccc3070 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", "dpmpp_3m","dpmpp_3m_sde", "ddim", "uni_pc", "uni_pc_bh2"]+ "dpmpp_2m", "dpmpp_2m_sde", "dpmpp_2m_sde_gpu", "dpmpp_3m", "dpmpp_3m_sde", "dpmpp_3m_sde_gpu", "ddim", "uni_pc", "uni_pc_bh2"]def __init__(self, model, steps, device, sampler=None, scheduler=None, denoise=None, model_options={}):self.model = model