stable-diffusion-webui/modules/sd_samplers_kdiffusion.py

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

325 lines
15 KiB
Python
Raw Normal View History

from collections import deque
import torch
2022-09-28 01:49:07 -06:00
import inspect
import k_diffusion.sampling
2023-08-08 09:35:31 -06:00
from modules import devices, sd_samplers_common, sd_samplers_extra, sd_samplers_cfg_denoiser
2023-08-05 20:37:25 -06:00
from modules.processing import StableDiffusionProcessing
from modules.shared import opts, state
import modules.shared as shared
2022-09-03 08:21:15 -06:00
samplers_k_diffusion = [
('Euler a', 'sample_euler_ancestral', ['k_euler_a', 'k_euler_ancestral'], {"uses_ensd": True}),
2022-10-06 05:12:52 -06:00
('Euler', 'sample_euler', ['k_euler'], {}),
('LMS', 'sample_lms', ['k_lms'], {}),
('Heun', 'sample_heun', ['k_heun'], {"second_order": True}),
('DPM2', 'sample_dpm_2', ['k_dpm_2'], {'discard_next_to_last_sigma': True}),
('DPM2 a', 'sample_dpm_2_ancestral', ['k_dpm_2_a'], {'discard_next_to_last_sigma': True, "uses_ensd": True}),
('DPM++ 2S a', 'sample_dpmpp_2s_ancestral', ['k_dpmpp_2s_a'], {"uses_ensd": True, "second_order": True}),
2022-11-05 09:32:22 -06:00
('DPM++ 2M', 'sample_dpmpp_2m', ['k_dpmpp_2m'], {}),
('DPM++ SDE', 'sample_dpmpp_sde', ['k_dpmpp_sde'], {"second_order": True, "brownian_noise": True}),
2023-05-22 11:06:57 -06:00
('DPM++ 2M SDE', 'sample_dpmpp_2m_sde', ['k_dpmpp_2m_sde_ka'], {"brownian_noise": True}),
('DPM fast', 'sample_dpm_fast', ['k_dpm_fast'], {"uses_ensd": True}),
('DPM adaptive', 'sample_dpm_adaptive', ['k_dpm_ad'], {"uses_ensd": True}),
2022-10-06 05:12:52 -06:00
('LMS Karras', 'sample_lms', ['k_lms_ka'], {'scheduler': 'karras'}),
('DPM2 Karras', 'sample_dpm_2', ['k_dpm_2_ka'], {'scheduler': 'karras', 'discard_next_to_last_sigma': True, "uses_ensd": True, "second_order": True}),
('DPM2 a Karras', 'sample_dpm_2_ancestral', ['k_dpm_2_a_ka'], {'scheduler': 'karras', 'discard_next_to_last_sigma': True, "uses_ensd": True, "second_order": True}),
('DPM++ 2S a Karras', 'sample_dpmpp_2s_ancestral', ['k_dpmpp_2s_a_ka'], {'scheduler': 'karras', "uses_ensd": True, "second_order": True}),
2022-11-05 09:32:22 -06:00
('DPM++ 2M Karras', 'sample_dpmpp_2m', ['k_dpmpp_2m_ka'], {'scheduler': 'karras'}),
('DPM++ SDE Karras', 'sample_dpmpp_sde', ['k_dpmpp_sde_ka'], {'scheduler': 'karras', "second_order": True, "brownian_noise": True}),
2023-05-22 11:06:57 -06:00
('DPM++ 2M SDE Karras', 'sample_dpmpp_2m_sde', ['k_dpmpp_2m_sde_ka'], {'scheduler': 'karras', "brownian_noise": True}),
('DPM++ 2M SDE Exponential', 'sample_dpmpp_2m_sde', ['k_dpmpp_2m_sde_exp'], {'scheduler': 'exponential', "brownian_noise": True}),
('Restart', sd_samplers_extra.restart_sampler, ['restart'], {'scheduler': 'karras'}),
2022-09-03 08:21:15 -06:00
]
2023-07-17 22:32:01 -06:00
2022-09-03 08:21:15 -06:00
samplers_data_k_diffusion = [
sd_samplers_common.SamplerData(label, lambda model, funcname=funcname: KDiffusionSampler(funcname, model), aliases, options)
2022-10-06 05:12:52 -06:00
for label, funcname, aliases, options in samplers_k_diffusion
if callable(funcname) or hasattr(k_diffusion.sampling, funcname)
2022-09-03 08:21:15 -06:00
]
sampler_extra_params = {
2022-09-28 01:49:07 -06:00
'sample_euler': ['s_churn', 's_tmin', 's_tmax', 's_noise'],
'sample_heun': ['s_churn', 's_tmin', 's_tmax', 's_noise'],
'sample_dpm_2': ['s_churn', 's_tmin', 's_tmax', 's_noise'],
}
k_diffusion_samplers_map = {x.name: x for x in samplers_data_k_diffusion}
2023-05-22 09:02:05 -06:00
k_diffusion_scheduler = {
2023-05-23 10:18:09 -06:00
'Automatic': None,
2023-05-22 09:02:05 -06:00
'karras': k_diffusion.sampling.get_sigmas_karras,
'exponential': k_diffusion.sampling.get_sigmas_exponential,
'polyexponential': k_diffusion.sampling.get_sigmas_polyexponential
}
class TorchHijack:
def __init__(self, sampler_noises):
# Using a deque to efficiently receive the sampler_noises in the same order as the previous index-based
# implementation.
self.sampler_noises = deque(sampler_noises)
def __getattr__(self, item):
if item == 'randn_like':
return self.randn_like
if hasattr(torch, item):
return getattr(torch, item)
raise AttributeError(f"'{type(self).__name__}' object has no attribute '{item}'")
def randn_like(self, x):
if self.sampler_noises:
noise = self.sampler_noises.popleft()
if noise.shape == x.shape:
return noise
return devices.randn_like(x)
class KDiffusionSampler:
def __init__(self, funcname, sd_model):
2022-11-26 06:10:46 -07:00
denoiser = k_diffusion.external.CompVisVDenoiser if sd_model.parameterization == "v" else k_diffusion.external.CompVisDenoiser
self.model_wrap = denoiser(sd_model, quantize=shared.opts.enable_quantization)
self.funcname = funcname
self.func = funcname if callable(funcname) else getattr(k_diffusion.sampling, self.funcname)
2022-09-28 01:49:07 -06:00
self.extra_params = sampler_extra_params.get(funcname, [])
2023-08-08 09:35:31 -06:00
self.model_wrap_cfg = sd_samplers_cfg_denoiser.CFGDenoiser(self.model_wrap)
self.sampler_noises = None
2022-09-19 07:42:56 -06:00
self.stop_at = None
self.eta = None
self.config = None # set by the function calling the constructor
self.last_latent = None
self.s_min_uncond = None
2023-08-05 20:37:25 -06:00
# NOTE: These are also defined in the StableDiffusionProcessing class.
# They should have been here to begin with but we're going to
# leave that class __init__ signature alone.
self.s_churn = 0.0
self.s_tmin = 0.0
self.s_tmax = float('inf')
self.s_noise = 1.0
self.conditioning_key = sd_model.model.conditioning_key
2022-09-06 10:33:51 -06:00
def callback_state(self, d):
step = d['i']
latent = d["denoised"]
if opts.live_preview_content == "Combined":
sd_samplers_common.store_latent(latent)
self.last_latent = latent
if self.stop_at is not None and step > self.stop_at:
raise sd_samplers_common.InterruptedException
state.sampling_step = step
shared.total_tqdm.update()
def launch_sampling(self, steps, func):
state.sampling_steps = steps
state.sampling_step = 0
try:
return func()
except RecursionError:
print(
'Encountered RecursionError during sampling, returning last latent. '
'rho >5 with a polyexponential scheduler may cause this error. '
'You should try to use a smaller rho value instead.'
)
return self.last_latent
except sd_samplers_common.InterruptedException:
return self.last_latent
2022-09-06 10:33:51 -06:00
def number_of_needed_noises(self, p):
return p.steps
2023-08-05 20:37:25 -06:00
def initialize(self, p: StableDiffusionProcessing):
2022-09-19 07:42:56 -06:00
self.model_wrap_cfg.mask = p.mask if hasattr(p, 'mask') else None
self.model_wrap_cfg.nmask = p.nmask if hasattr(p, 'nmask') else None
self.model_wrap_cfg.step = 0
self.model_wrap_cfg.image_cfg_scale = getattr(p, 'image_cfg_scale', None)
self.eta = p.eta if p.eta is not None else opts.eta_ancestral
2023-03-28 16:18:28 -06:00
self.s_min_uncond = getattr(p, 's_min_uncond', 0.0)
k_diffusion.sampling.torch = TorchHijack(self.sampler_noises if self.sampler_noises is not None else [])
extra_params_kwargs = {}
2022-09-28 01:49:07 -06:00
for param_name in self.extra_params:
if hasattr(p, param_name) and param_name in inspect.signature(self.func).parameters:
extra_params_kwargs[param_name] = getattr(p, param_name)
if 'eta' in inspect.signature(self.func).parameters:
if self.eta != 1.0:
p.extra_generation_params["Eta"] = self.eta
extra_params_kwargs['eta'] = self.eta
2023-08-05 20:37:25 -06:00
if len(self.extra_params) > 0:
2023-08-05 21:42:20 -06:00
s_churn = getattr(opts, 's_churn', p.s_churn)
s_tmin = getattr(opts, 's_tmin', p.s_tmin)
2023-08-05 21:50:33 -06:00
s_tmax = getattr(opts, 's_tmax', p.s_tmax) or self.s_tmax # 0 = inf
2023-08-05 21:42:20 -06:00
s_noise = getattr(opts, 's_noise', p.s_noise)
2023-08-05 20:37:25 -06:00
if s_churn != self.s_churn:
extra_params_kwargs['s_churn'] = s_churn
p.s_churn = s_churn
p.extra_generation_params['Sigma churn'] = s_churn
if s_tmin != self.s_tmin:
2023-08-05 21:42:20 -06:00
extra_params_kwargs['s_tmin'] = s_tmin
2023-08-05 20:37:25 -06:00
p.s_tmin = s_tmin
p.extra_generation_params['Sigma tmin'] = s_tmin
if s_tmax != self.s_tmax:
2023-08-05 21:42:20 -06:00
extra_params_kwargs['s_tmax'] = s_tmax
2023-08-05 20:37:25 -06:00
p.s_tmax = s_tmax
p.extra_generation_params['Sigma tmax'] = s_tmax
if s_noise != self.s_noise:
2023-08-05 21:42:20 -06:00
extra_params_kwargs['s_noise'] = s_noise
2023-08-05 20:37:25 -06:00
p.s_noise = s_noise
p.extra_generation_params['Sigma noise'] = s_noise
return extra_params_kwargs
def get_sigmas(self, p, steps):
discard_next_to_last_sigma = self.config is not None and self.config.options.get('discard_next_to_last_sigma', False)
if opts.always_discard_next_to_last_sigma and not discard_next_to_last_sigma:
discard_next_to_last_sigma = True
p.extra_generation_params["Discard penultimate sigma"] = True
steps += 1 if discard_next_to_last_sigma else 0
2022-12-26 13:49:13 -07:00
if p.sampler_noise_scheduler_override:
2022-10-06 14:27:01 -06:00
sigmas = p.sampler_noise_scheduler_override(steps)
2023-05-23 10:18:09 -06:00
elif opts.k_sched_type != "Automatic":
m_sigma_min, m_sigma_max = (self.model_wrap.sigmas[0].item(), self.model_wrap.sigmas[-1].item())
sigma_min, sigma_max = (0.1, 10) if opts.use_old_karras_scheduler_sigmas else (m_sigma_min, m_sigma_max)
2023-05-22 09:02:05 -06:00
sigmas_kwargs = {
'sigma_min': sigma_min,
'sigma_max': sigma_max,
2023-05-22 09:02:05 -06:00
}
sigmas_func = k_diffusion_scheduler[opts.k_sched_type]
p.extra_generation_params["Schedule type"] = opts.k_sched_type
if opts.sigma_min != m_sigma_min and opts.sigma_min != 0:
sigmas_kwargs['sigma_min'] = opts.sigma_min
p.extra_generation_params["Schedule min sigma"] = opts.sigma_min
if opts.sigma_max != m_sigma_max and opts.sigma_max != 0:
sigmas_kwargs['sigma_max'] = opts.sigma_max
p.extra_generation_params["Schedule max sigma"] = opts.sigma_max
default_rho = 1. if opts.k_sched_type == "polyexponential" else 7.
if opts.k_sched_type != 'exponential' and opts.rho != 0 and opts.rho != default_rho:
2023-05-22 21:34:51 -06:00
sigmas_kwargs['rho'] = opts.rho
p.extra_generation_params["Schedule rho"] = opts.rho
2023-05-22 09:02:05 -06:00
sigmas = sigmas_func(n=steps, **sigmas_kwargs, device=shared.device)
2022-10-06 14:27:01 -06:00
elif self.config is not None and self.config.options.get('scheduler', None) == 'karras':
sigma_min, sigma_max = (0.1, 10) if opts.use_old_karras_scheduler_sigmas else (self.model_wrap.sigmas[0].item(), self.model_wrap.sigmas[-1].item())
sigmas = k_diffusion.sampling.get_sigmas_karras(n=steps, sigma_min=sigma_min, sigma_max=sigma_max, device=shared.device)
elif self.config is not None and self.config.options.get('scheduler', None) == 'exponential':
m_sigma_min, m_sigma_max = (self.model_wrap.sigmas[0].item(), self.model_wrap.sigmas[-1].item())
sigmas = k_diffusion.sampling.get_sigmas_exponential(n=steps, sigma_min=m_sigma_min, sigma_max=m_sigma_max, device=shared.device)
else:
2022-10-06 14:27:01 -06:00
sigmas = self.model_wrap.get_sigmas(steps)
if discard_next_to_last_sigma:
2022-12-18 20:16:42 -07:00
sigmas = torch.cat([sigmas[:-2], sigmas[-1:]])
return sigmas
2023-02-15 01:57:18 -07:00
def create_noise_sampler(self, x, sigmas, p):
"""For DPM++ SDE: manually create noise sampler to enable deterministic results across different batch sizes"""
if shared.opts.no_dpmpp_sde_batch_determinism:
return None
from k_diffusion.sampling import BrownianTreeNoiseSampler
sigma_min, sigma_max = sigmas[sigmas > 0].min(), sigmas.max()
2023-02-15 01:57:18 -07:00
current_iter_seeds = p.all_seeds[p.iteration * p.batch_size:(p.iteration + 1) * p.batch_size]
return BrownianTreeNoiseSampler(x, sigma_min, sigma_max, seed=current_iter_seeds)
def sample_img2img(self, p, x, noise, conditioning, unconditional_conditioning, steps=None, image_conditioning=None):
steps, t_enc = sd_samplers_common.setup_img2img_steps(p, steps)
sigmas = self.get_sigmas(p, steps)
sigma_sched = sigmas[steps - t_enc - 1:]
xi = x + noise * sigma_sched[0]
extra_params_kwargs = self.initialize(p)
parameters = inspect.signature(self.func).parameters
if 'sigma_min' in parameters:
2022-10-10 17:36:00 -06:00
## last sigma is zero which isn't allowed by DPM Fast & Adaptive so taking value before last
extra_params_kwargs['sigma_min'] = sigma_sched[-2]
if 'sigma_max' in parameters:
extra_params_kwargs['sigma_max'] = sigma_sched[0]
if 'n' in parameters:
extra_params_kwargs['n'] = len(sigma_sched) - 1
if 'sigma_sched' in parameters:
extra_params_kwargs['sigma_sched'] = sigma_sched
if 'sigmas' in parameters:
extra_params_kwargs['sigmas'] = sigma_sched
if self.config.options.get('brownian_noise', False):
2023-02-15 01:57:18 -07:00
noise_sampler = self.create_noise_sampler(x, sigmas, p)
extra_params_kwargs['noise_sampler'] = noise_sampler
self.model_wrap_cfg.init_latent = x
self.last_latent = x
extra_args = {
'cond': conditioning,
'image_cond': image_conditioning,
'uncond': unconditional_conditioning,
'cond_scale': p.cfg_scale,
2023-03-28 16:18:28 -06:00
's_min_uncond': self.s_min_uncond
}
samples = self.launch_sampling(t_enc + 1, lambda: self.func(self.model_wrap_cfg, xi, extra_args=extra_args, disable=False, callback=self.callback_state, **extra_params_kwargs))
if self.model_wrap_cfg.padded_cond_uncond:
p.extra_generation_params["Pad conds"] = True
return samples
def sample(self, p, x, conditioning, unconditional_conditioning, steps=None, image_conditioning=None):
2022-09-19 07:42:56 -06:00
steps = steps or p.steps
sigmas = self.get_sigmas(p, steps)
2022-10-06 05:12:52 -06:00
x = x * sigmas[0]
extra_params_kwargs = self.initialize(p)
parameters = inspect.signature(self.func).parameters
if 'sigma_min' in parameters:
2022-09-29 04:30:33 -06:00
extra_params_kwargs['sigma_min'] = self.model_wrap.sigmas[0].item()
extra_params_kwargs['sigma_max'] = self.model_wrap.sigmas[-1].item()
if 'n' in parameters:
2022-09-29 04:30:33 -06:00
extra_params_kwargs['n'] = steps
else:
extra_params_kwargs['sigmas'] = sigmas
if self.config.options.get('brownian_noise', False):
2023-02-15 01:57:18 -07:00
noise_sampler = self.create_noise_sampler(x, sigmas, p)
extra_params_kwargs['noise_sampler'] = noise_sampler
self.last_latent = x
samples = self.launch_sampling(steps, lambda: self.func(self.model_wrap_cfg, x, extra_args={
'cond': conditioning,
'image_cond': image_conditioning,
'uncond': unconditional_conditioning,
2023-03-28 16:18:28 -06:00
'cond_scale': p.cfg_scale,
's_min_uncond': self.s_min_uncond
}, disable=False, callback=self.callback_state, **extra_params_kwargs))
if self.model_wrap_cfg.padded_cond_uncond:
p.extra_generation_params["Pad conds"] = True
2022-09-19 07:42:56 -06:00
return samples