Merge 6fd91a8f006636bf0a2ddf306a280b6ca0591af7 into 97ae6ef46001ed444ef1f2bd085beae98c17c240

This commit is contained in:
JettHu 2024-06-20 00:30:29 +08:00 committed by GitHub
commit 6033118170
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
5 changed files with 62 additions and 6 deletions

View File

@ -841,3 +841,40 @@ def sample_heunpp2(model, x, sigmas, extra_args=None, callback=None, disable=Non
d_prime = w1 * d + w2 * d_2 + w3 * d_3
x = x + d_prime * dt
return x
@torch.no_grad()
def sample_tcd(
model,
x,
sigmas,
extra_args=None,
callback=None,
disable=None,
noise_sampler=None,
eta=0.3,
):
extra_args = {} if extra_args is None else extra_args
noise_sampler = default_noise_sampler(x) if noise_sampler is None else noise_sampler
s_in = x.new_ones([x.shape[0]])
model_sampling = model.inner_model.model_patcher.get_model_object("model_sampling")
timesteps_s = torch.floor((1 - eta) * model_sampling.timestep(sigmas)).to(dtype=torch.long).detach().cpu()
timesteps_s[-1] = 0
alpha_prod_s = model_sampling.alphas_cumprod[timesteps_s]
beta_prod_s = 1 - alpha_prod_s
for i in trange(len(sigmas) - 1, disable=disable):
denoised = model(x, sigmas[i] * s_in, **extra_args) # predicted_original_sample
eps = (x - denoised) / sigmas[i]
denoised = alpha_prod_s[i + 1].sqrt() * denoised + beta_prod_s[i + 1].sqrt() * eps
if callback is not None:
callback({"x": x, "i": i, "sigma": sigmas[i], "sigma_hat": sigmas[i], "denoised": denoised})
x = denoised
if eta > 0 and sigmas[i + 1] > 0:
noise = noise_sampler(sigmas[i], sigmas[i + 1])
x = x / alpha_prod_s[i+1].sqrt() + noise * (sigmas[i+1]**2 + 1 - 1/alpha_prod_s[i+1]).sqrt()
else:
x *= torch.sqrt(1.0 + sigmas[i + 1] ** 2)
return x

View File

@ -77,12 +77,10 @@ class ModelSamplingDiscrete(torch.nn.Module):
self.linear_start = linear_start
self.linear_end = linear_end
# self.register_buffer('betas', torch.tensor(betas, dtype=torch.float32))
# self.register_buffer('alphas_cumprod', torch.tensor(alphas_cumprod, dtype=torch.float32))
# self.register_buffer('alphas_cumprod_prev', torch.tensor(alphas_cumprod_prev, dtype=torch.float32))
sigmas = ((1 - alphas_cumprod) / alphas_cumprod) ** 0.5
self.set_sigmas(sigmas)
# register alphas_cumprod for some sampler, such as tcd.
self.register_buffer("alphas_cumprod", alphas_cumprod.float())
def set_sigmas(self, sigmas):
self.register_buffer('sigmas', sigmas.float())

View File

@ -539,7 +539,7 @@ class Sampler:
KSAMPLER_NAMES = ["euler", "euler_ancestral", "heun", "heunpp2","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_sde", "dpmpp_3m_sde_gpu", "ddpm", "lcm"]
"dpmpp_2m", "dpmpp_2m_sde", "dpmpp_2m_sde_gpu", "dpmpp_3m_sde", "dpmpp_3m_sde_gpu", "ddpm", "lcm", "tcd"]
class KSAMPLER(Sampler):
def __init__(self, sampler_function, extra_options={}, inpaint_options={}):

View File

@ -336,6 +336,23 @@ class SamplerDPMAdaptative:
"s_noise":s_noise })
return (sampler, )
class SamplerTCD:
@classmethod
def INPUT_TYPES(s):
return {
"required": {
"eta": ("FLOAT", {"default": 0.3, "min": 0.0, "max": 1.0, "step": 0.01}),
}
}
RETURN_TYPES = ("SAMPLER",)
CATEGORY = "sampling/custom_sampling/samplers"
FUNCTION = "get_sampler"
def get_sampler(self, eta=0.3):
sampler = comfy.samplers.ksampler("tcd", {"eta": eta})
return (sampler, )
class Noise_EmptyNoise:
def __init__(self):
self.seed = 0
@ -627,6 +644,7 @@ NODE_CLASS_MAPPINGS = {
"SamplerDPMPP_2M_SDE": SamplerDPMPP_2M_SDE,
"SamplerDPMPP_SDE": SamplerDPMPP_SDE,
"SamplerDPMAdaptative": SamplerDPMAdaptative,
"SamplerTCD": SamplerTCD,
"SplitSigmas": SplitSigmas,
"SplitSigmasDenoise": SplitSigmasDenoise,
"FlipSigmas": FlipSigmas,

View File

@ -73,7 +73,7 @@ class ModelSamplingDiscrete:
@classmethod
def INPUT_TYPES(s):
return {"required": { "model": ("MODEL",),
"sampling": (["eps", "v_prediction", "lcm", "x0"],),
"sampling": (["eps", "v_prediction", "lcm", "x0", "tcd"],),
"zsnr": ("BOOLEAN", {"default": False}),
}}
@ -95,6 +95,9 @@ class ModelSamplingDiscrete:
sampling_base = ModelSamplingDiscreteDistilled
elif sampling == "x0":
sampling_type = X0
elif sampling == "tcd":
sampling_type = comfy.model_sampling.EPS
sampling_base= ModelSamplingDiscreteDistilled
class ModelSamplingAdvanced(sampling_base, sampling_type):
pass