Add DPM-Solver++(2M) SDE and exponential scheduler.

exponential scheduler is the one recommended with this sampler.

comfy/k_diffusion/sampling.py

@@ -605,3 +605,46 @@ def sample_dpmpp_2m(model, x, sigmas, extra_args=None, callback=None, disable=No
             x = (sigma_fn(t_next) / sigma_fn(t)) * x - (-h).expm1() * denoised_d
         old_denoised = denoised
     return x
+
+@torch.no_grad()
+def sample_dpmpp_2m_sde(model, x, sigmas, extra_args=None, callback=None, disable=None, eta=1., s_noise=1., noise_sampler=None, solver_type='midpoint'):
+    """DPM-Solver++(2M) SDE."""
+
+    if solver_type not in {'heun', 'midpoint'}:
+        raise ValueError('solver_type must be \'heun\' or \'midpoint\'')
+
+    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]])
+
+    old_denoised = None
+    h_last = 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:
+            # DPM-Solver++(2M) SDE
+            t, s = -sigmas[i].log(), -sigmas[i + 1].log()
+            h = s - t
+            eta_h = eta * h
+
+            x = sigmas[i + 1] / sigmas[i] * (-eta_h).exp() * x + (-h - eta_h).expm1().neg() * denoised
+
+            if old_denoised is not None:
+                r = h_last / h
+                if solver_type == 'heun':
+                    x = x + ((-h - eta_h).expm1().neg() / (-h - eta_h) + 1) * (1 / r) * (denoised - old_denoised)
+                elif 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
+
+        old_denoised = denoised
+        h_last = h
+    return x

comfy/samplers.py

@@ -495,10 +495,10 @@ def encode_adm(noise_augmentor, conds, batch_size, device):
 
 
 class KSampler:
-    SCHEDULERS = ["normal", "karras", "simple", "ddim_uniform"]
+    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_2m", "ddim", "uni_pc", "uni_pc_bh2"]
+                "dpmpp_2m", "dpmpp_2m_sde", "ddim", "uni_pc", "uni_pc_bh2"]
 
     def __init__(self, model, steps, device, sampler=None, scheduler=None, denoise=None, model_options={}):
         self.model = model
@@ -532,6 +532,8 @@ class KSampler:
 
         if self.scheduler == "karras":
             sigmas = k_diffusion_sampling.get_sigmas_karras(n=steps, sigma_min=self.sigma_min, sigma_max=self.sigma_max)
+        elif self.scheduler == "exponential":
+            sigmas = k_diffusion_sampling.get_sigmas_exponential(n=steps, sigma_min=self.sigma_min, sigma_max=self.sigma_max)
         elif self.scheduler == "normal":
             sigmas = self.model_wrap.get_sigmas(steps)
         elif self.scheduler == "simple":