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.

434 lines
21 KiB
Python
Raw Normal View History

from collections import deque
import torch
2022-09-28 15:49:07 +08:00
import inspect
import k_diffusion.sampling
from modules import prompt_parser, devices, sd_samplers_common
from modules.shared import opts, state
import modules.shared as shared
from modules.script_callbacks import CFGDenoiserParams, cfg_denoiser_callback
2023-02-11 10:18:38 +08:00
from modules.script_callbacks import CFGDenoisedParams, cfg_denoised_callback
from modules.script_callbacks import AfterCFGCallbackParams, cfg_after_cfg_callback
2022-09-03 22:21:15 +08:00
samplers_k_diffusion = [
('Euler a', 'sample_euler_ancestral', ['k_euler_a', 'k_euler_ancestral'], {"uses_ensd": True}),
2022-10-06 19:12:52 +08:00
('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 23:32:22 +08:00
('DPM++ 2M', 'sample_dpmpp_2m', ['k_dpmpp_2m'], {}),
('DPM++ SDE', 'sample_dpmpp_sde', ['k_dpmpp_sde'], {"second_order": True, "brownian_noise": True}),
('DPM++ 2M SDE', 'sample_dpmpp_2m_sde', ['k_dpmpp_2m_sde_ka'], {"brownian_noise": True, 'discard_next_to_last_sigma': 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 19:12:52 +08:00
('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 23:32:22 +08:00
('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}),
('DPM++ 2M SDE Karras', 'sample_dpmpp_2m_sde', ['k_dpmpp_2m_sde_ka'], {'scheduler': 'karras', "brownian_noise": True, 'discard_next_to_last_sigma': True}),
2022-09-03 22:21:15 +08:00
]
samplers_data_k_diffusion = [
sd_samplers_common.SamplerData(label, lambda model, funcname=funcname: KDiffusionSampler(funcname, model), aliases, options)
2022-10-06 19:12:52 +08:00
for label, funcname, aliases, options in samplers_k_diffusion
2022-09-03 22:21:15 +08:00
if hasattr(k_diffusion.sampling, funcname)
]
sampler_extra_params = {
2022-09-28 15:49:07 +08:00
'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 23:02:05 +08:00
k_diffusion_scheduler = {
2023-05-23 23:48:23 +08:00
'None': None,
2023-05-22 23:02:05 +08:00
'karras': k_diffusion.sampling.get_sigmas_karras,
'exponential': k_diffusion.sampling.get_sigmas_exponential,
'polyexponential': k_diffusion.sampling.get_sigmas_polyexponential
}
class CFGDenoiser(torch.nn.Module):
"""
Classifier free guidance denoiser. A wrapper for stable diffusion model (specifically for unet)
that can take a noisy picture and produce a noise-free picture using two guidances (prompts)
instead of one. Originally, the second prompt is just an empty string, but we use non-empty
negative prompt.
"""
def __init__(self, model):
super().__init__()
self.inner_model = model
self.mask = None
self.nmask = None
self.init_latent = None
2022-09-15 18:10:16 +08:00
self.step = 0
self.image_cfg_scale = None
def combine_denoised(self, x_out, conds_list, uncond, cond_scale):
denoised_uncond = x_out[-uncond.shape[0]:]
denoised = torch.clone(denoised_uncond)
for i, conds in enumerate(conds_list):
for cond_index, weight in conds:
denoised[i] += (x_out[cond_index] - denoised_uncond[i]) * (weight * cond_scale)
return denoised
def combine_denoised_for_edit_model(self, x_out, cond_scale):
out_cond, out_img_cond, out_uncond = x_out.chunk(3)
denoised = out_uncond + cond_scale * (out_cond - out_img_cond) + self.image_cfg_scale * (out_img_cond - out_uncond)
return denoised
2023-03-29 06:18:28 +08:00
def forward(self, x, sigma, uncond, cond, cond_scale, s_min_uncond, image_cond):
if state.interrupted or state.skipped:
raise sd_samplers_common.InterruptedException
# at self.image_cfg_scale == 1.0 produced results for edit model are the same as with normal sampling,
# so is_edit_model is set to False to support AND composition.
is_edit_model = shared.sd_model.cond_stage_key == "edit" and self.image_cfg_scale is not None and self.image_cfg_scale != 1.0
conds_list, tensor = prompt_parser.reconstruct_multicond_batch(cond, self.step)
2022-09-15 18:10:16 +08:00
uncond = prompt_parser.reconstruct_cond_batch(uncond, self.step)
2023-05-10 16:05:02 +08:00
assert not is_edit_model or all(len(conds) == 1 for conds in conds_list), "AND is not supported for InstructPix2Pix checkpoint (unless using Image CFG scale = 1.0)"
batch_size = len(conds_list)
repeats = [len(conds_list[i]) for i in range(batch_size)]
if shared.sd_model.model.conditioning_key == "crossattn-adm":
image_uncond = torch.zeros_like(image_cond)
make_condition_dict = lambda c_crossattn, c_adm: {"c_crossattn": c_crossattn, "c_adm": c_adm}
else:
image_uncond = image_cond
make_condition_dict = lambda c_crossattn, c_concat: {"c_crossattn": c_crossattn, "c_concat": [c_concat]}
if not is_edit_model:
x_in = torch.cat([torch.stack([x[i] for _ in range(n)]) for i, n in enumerate(repeats)] + [x])
sigma_in = torch.cat([torch.stack([sigma[i] for _ in range(n)]) for i, n in enumerate(repeats)] + [sigma])
image_cond_in = torch.cat([torch.stack([image_cond[i] for _ in range(n)]) for i, n in enumerate(repeats)] + [image_uncond])
else:
x_in = torch.cat([torch.stack([x[i] for _ in range(n)]) for i, n in enumerate(repeats)] + [x] + [x])
sigma_in = torch.cat([torch.stack([sigma[i] for _ in range(n)]) for i, n in enumerate(repeats)] + [sigma] + [sigma])
image_cond_in = torch.cat([torch.stack([image_cond[i] for _ in range(n)]) for i, n in enumerate(repeats)] + [image_uncond] + [torch.zeros_like(self.init_latent)])
2023-02-24 13:04:23 +08:00
denoiser_params = CFGDenoiserParams(x_in, image_cond_in, sigma_in, state.sampling_step, state.sampling_steps, tensor, uncond)
cfg_denoiser_callback(denoiser_params)
x_in = denoiser_params.x
image_cond_in = denoiser_params.image_cond
sigma_in = denoiser_params.sigma
tensor = denoiser_params.text_cond
uncond = denoiser_params.text_uncond
skip_uncond = False
2022-10-31 07:48:33 +08:00
# alternating uncond allows for higher thresholds without the quality loss normally expected from raising it
if self.step % 2 and s_min_uncond > 0 and sigma[0] < s_min_uncond and not is_edit_model:
skip_uncond = True
x_in = x_in[:-batch_size]
sigma_in = sigma_in[:-batch_size]
2023-03-29 06:18:28 +08:00
# TODO add infotext entry
if shared.opts.pad_cond_uncond and tensor.shape[1] != uncond.shape[1]:
empty = shared.sd_model.cond_stage_model_empty_prompt
num_repeats = (tensor.shape[1] - uncond.shape[1]) // empty.shape[1]
if num_repeats < 0:
tensor = torch.cat([tensor, empty.repeat((tensor.shape[0], -num_repeats, 1))], axis=1)
elif num_repeats > 0:
uncond = torch.cat([uncond, empty.repeat((uncond.shape[0], num_repeats, 1))], axis=1)
if tensor.shape[1] == uncond.shape[1] or skip_uncond:
if is_edit_model:
cond_in = torch.cat([tensor, uncond, uncond])
elif skip_uncond:
cond_in = tensor
else:
cond_in = torch.cat([tensor, uncond])
if shared.batch_cond_uncond:
x_out = self.inner_model(x_in, sigma_in, cond=make_condition_dict([cond_in], image_cond_in))
else:
x_out = torch.zeros_like(x_in)
for batch_offset in range(0, x_out.shape[0], batch_size):
a = batch_offset
b = a + batch_size
x_out[a:b] = self.inner_model(x_in[a:b], sigma_in[a:b], cond=make_condition_dict([cond_in[a:b]], image_cond_in[a:b]))
else:
x_out = torch.zeros_like(x_in)
batch_size = batch_size*2 if shared.batch_cond_uncond else batch_size
for batch_offset in range(0, tensor.shape[0], batch_size):
a = batch_offset
b = min(a + batch_size, tensor.shape[0])
if not is_edit_model:
c_crossattn = [tensor[a:b]]
else:
c_crossattn = torch.cat([tensor[a:b]], uncond)
x_out[a:b] = self.inner_model(x_in[a:b], sigma_in[a:b], cond=make_condition_dict(c_crossattn, image_cond_in[a:b]))
if not skip_uncond:
2023-03-29 06:18:28 +08:00
x_out[-uncond.shape[0]:] = self.inner_model(x_in[-uncond.shape[0]:], sigma_in[-uncond.shape[0]:], cond=make_condition_dict([uncond], image_cond_in[-uncond.shape[0]:]))
denoised_image_indexes = [x[0][0] for x in conds_list]
if skip_uncond:
fake_uncond = torch.cat([x_out[i:i+1] for i in denoised_image_indexes])
x_out = torch.cat([x_out, fake_uncond]) # we skipped uncond denoising, so we put cond-denoised image to where the uncond-denoised image should be
denoised_params = CFGDenoisedParams(x_out, state.sampling_step, state.sampling_steps, self.inner_model)
2023-02-11 10:18:38 +08:00
cfg_denoised_callback(denoised_params)
devices.test_for_nans(x_out, "unet")
if opts.live_preview_content == "Prompt":
sd_samplers_common.store_latent(torch.cat([x_out[i:i+1] for i in denoised_image_indexes]))
elif opts.live_preview_content == "Negative prompt":
sd_samplers_common.store_latent(x_out[-uncond.shape[0]:])
if is_edit_model:
denoised = self.combine_denoised_for_edit_model(x_out, cond_scale)
elif skip_uncond:
denoised = self.combine_denoised(x_out, conds_list, uncond, 1.0)
else:
denoised = self.combine_denoised(x_out, conds_list, uncond, cond_scale)
if self.mask is not None:
denoised = self.init_latent * self.mask + self.nmask * denoised
after_cfg_callback_params = AfterCFGCallbackParams(denoised, state.sampling_step, state.sampling_steps)
cfg_after_cfg_callback(after_cfg_callback_params)
denoised = after_cfg_callback_params.x
2022-09-15 18:10:16 +08:00
self.step += 1
return denoised
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
if opts.randn_source == "CPU" or x.device.type == 'mps':
return torch.randn_like(x, device=devices.cpu).to(x.device)
else:
return torch.randn_like(x)
class KDiffusionSampler:
def __init__(self, funcname, sd_model):
2022-11-26 21:10:46 +08:00
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 = getattr(k_diffusion.sampling, self.funcname)
2022-09-28 15:49:07 +08:00
self.extra_params = sampler_extra_params.get(funcname, [])
self.model_wrap_cfg = CFGDenoiser(self.model_wrap)
self.sampler_noises = None
2022-09-19 21:42:56 +08:00
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
self.conditioning_key = sd_model.model.conditioning_key
2022-09-07 00:33:51 +08:00
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
2022-09-07 00:33:51 +08:00
def number_of_needed_noises(self, p):
return p.steps
def initialize(self, p):
2022-09-19 21:42:56 +08:00
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
2023-03-29 06:18:28 +08:00
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 [])
2023-05-23 23:48:23 +08:00
if opts.k_sched_type != "None":
2023-05-23 11:34:51 +08:00
p.extra_generation_params["KDiffusion Scheduler Type"] = opts.k_sched_type
p.extra_generation_params["KDiffusion Scheduler sigma_max"] = opts.sigma_max
p.extra_generation_params["KDiffusion Scheduler sigma_min"] = opts.sigma_min
p.extra_generation_params["KDiffusion Scheduler rho"] = opts.rho
extra_params_kwargs = {}
2022-09-28 15:49:07 +08:00
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
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
2022-12-27 04:49:13 +08:00
if p.sampler_noise_scheduler_override:
2022-10-07 04:27:01 +08:00
sigmas = p.sampler_noise_scheduler_override(steps)
2023-05-23 23:48:23 +08:00
elif opts.k_sched_type != "None":
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())
2023-05-23 11:34:51 +08:00
sigmas_func = k_diffusion_scheduler[opts.k_sched_type]
2023-05-22 23:02:05 +08:00
sigmas_kwargs = {
2023-05-23 11:34:51 +08:00
'sigma_min': opts.sigma_min or sigma_min,
'sigma_max': opts.sigma_max or sigma_max
2023-05-22 23:02:05 +08:00
}
2023-05-23 11:34:51 +08:00
if opts.k_sched_type != 'exponential':
sigmas_kwargs['rho'] = opts.rho
2023-05-22 23:02:05 +08:00
sigmas = sigmas_func(n=steps, **sigmas_kwargs, device=shared.device)
2022-10-07 04:27:01 +08:00
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)
else:
2022-10-07 04:27:01 +08:00
sigmas = self.model_wrap.get_sigmas(steps)
if discard_next_to_last_sigma:
2022-12-19 11:16:42 +08:00
sigmas = torch.cat([sigmas[:-2], sigmas[-1:]])
return sigmas
2023-02-15 16:57:18 +08:00
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 16:57:18 +08:00
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:
2022-10-11 07:36:00 +08:00
## 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 16:57:18 +08:00
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,
2023-03-29 06:18:28 +08:00
'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))
return samples
def sample(self, p, x, conditioning, unconditional_conditioning, steps=None, image_conditioning=None):
2022-09-19 21:42:56 +08:00
steps = steps or p.steps
sigmas = self.get_sigmas(p, steps)
2022-10-06 19:12:52 +08:00
x = x * sigmas[0]
extra_params_kwargs = self.initialize(p)
parameters = inspect.signature(self.func).parameters
if 'sigma_min' in parameters:
2022-09-29 18:30:33 +08:00
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 18:30:33 +08:00
extra_params_kwargs['n'] = steps
else:
extra_params_kwargs['sigmas'] = sigmas
if self.config.options.get('brownian_noise', False):
2023-02-15 16:57:18 +08:00
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,
2023-03-29 06:18:28 +08:00
'cond_scale': p.cfg_scale,
's_min_uncond': self.s_min_uncond
}, disable=False, callback=self.callback_state, **extra_params_kwargs))
2022-09-19 21:42:56 +08:00
return samples