Spaces:
Runtime error
Runtime error
File size: 17,994 Bytes
424f073 5be29af 424f073 5be29af 424f073 5be29af 424f073 5be29af 424f073 5be29af 424f073 32cf151 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 |
# Code is based on ReNoise https://github.com/garibida/ReNoise-Inversion
from diffusers import EulerAncestralDiscreteScheduler
from diffusers.utils import BaseOutput
import torch
from typing import List, Optional, Tuple, Union
import numpy as np
from src.eunms import Epsilon_Update_Type
class EulerAncestralDiscreteSchedulerOutput(BaseOutput):
"""
Output class for the scheduler's `step` function output.
Args:
prev_sample (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)` for images):
Computed sample `(x_{t-1})` of previous timestep. `prev_sample` should be used as next model input in the
denoising loop.
pred_original_sample (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)` for images):
The predicted denoised sample `(x_{0})` based on the model output from the current timestep.
`pred_original_sample` can be used to preview progress or for guidance.
"""
prev_sample: torch.FloatTensor
pred_original_sample: Optional[torch.FloatTensor] = None
class MyEulerAncestralDiscreteScheduler(EulerAncestralDiscreteScheduler):
def set_noise_list(self, noise_list):
self.noise_list = noise_list
def set_noise_list_device(self, device):
if self.noise_list[0].device == device:
return
for i in range(len(self.noise_list)):
self.noise_list[i] = self.noise_list[i].to(device)
def get_noise_to_remove(self):
sigma_from = self.sigmas[self.step_index]
sigma_to = self.sigmas[self.step_index + 1]
sigma_up = (sigma_to**2 * (sigma_from**2 - sigma_to**2) / sigma_from**2) ** 0.5
return self.noise_list[self.step_index] * sigma_up\
def scale_model_input(
self, sample: torch.FloatTensor, timestep: Union[float, torch.FloatTensor]
) -> torch.FloatTensor:
"""
Ensures interchangeability with schedulers that need to scale the denoising model input depending on the
current timestep. Scales the denoising model input by `(sigma**2 + 1) ** 0.5` to match the Euler algorithm.
Args:
sample (`torch.FloatTensor`):
The input sample.
timestep (`int`, *optional*):
The current timestep in the diffusion chain.
Returns:
`torch.FloatTensor`:
A scaled input sample.
"""
self._init_step_index(timestep.view((1)))
return EulerAncestralDiscreteScheduler.scale_model_input(self, sample, timestep)
def step(
self,
model_output: torch.FloatTensor,
timestep: Union[float, torch.FloatTensor],
sample: torch.FloatTensor,
generator: Optional[torch.Generator] = None,
return_dict: bool = True,
) -> Union[EulerAncestralDiscreteSchedulerOutput, Tuple]:
"""
Predict the sample from the previous timestep by reversing the SDE. This function propagates the diffusion
process from the learned model outputs (most often the predicted noise).
Args:
model_output (`torch.FloatTensor`):
The direct output from learned diffusion model.
timestep (`float`):
The current discrete timestep in the diffusion chain.
sample (`torch.FloatTensor`):
A current instance of a sample created by the diffusion process.
generator (`torch.Generator`, *optional*):
A random number generator.
return_dict (`bool`):
Whether or not to return a
[`~schedulers.scheduling_euler_ancestral_discrete.EulerAncestralDiscreteSchedulerOutput`] or tuple.
Returns:
[`~schedulers.scheduling_euler_ancestral_discrete.EulerAncestralDiscreteSchedulerOutput`] or `tuple`:
If return_dict is `True`,
[`~schedulers.scheduling_euler_ancestral_discrete.EulerAncestralDiscreteSchedulerOutput`] is returned,
otherwise a tuple is returned where the first element is the sample tensor.
"""
if (
isinstance(timestep, int)
or isinstance(timestep, torch.IntTensor)
or isinstance(timestep, torch.LongTensor)
):
raise ValueError(
(
"Passing integer indices (e.g. from `enumerate(timesteps)`) as timesteps to"
" `EulerDiscreteScheduler.step()` is not supported. Make sure to pass"
" one of the `scheduler.timesteps` as a timestep."
),
)
if not self.is_scale_input_called:
logger.warning(
"The `scale_model_input` function should be called before `step` to ensure correct denoising. "
"See `StableDiffusionPipeline` for a usage example."
)
self._init_step_index(timestep.view((1)))
sigma = self.sigmas[self.step_index]
# Upcast to avoid precision issues when computing prev_sample
sample = sample.to(torch.float32)
# 1. compute predicted original sample (x_0) from sigma-scaled predicted noise
if self.config.prediction_type == "epsilon":
pred_original_sample = sample - sigma * model_output
elif self.config.prediction_type == "v_prediction":
# * c_out + input * c_skip
pred_original_sample = model_output * (-sigma / (sigma**2 + 1) ** 0.5) + (sample / (sigma**2 + 1))
elif self.config.prediction_type == "sample":
raise NotImplementedError("prediction_type not implemented yet: sample")
else:
raise ValueError(
f"prediction_type given as {self.config.prediction_type} must be one of `epsilon`, or `v_prediction`"
)
sigma_from = self.sigmas[self.step_index]
sigma_to = self.sigmas[self.step_index + 1]
sigma_up = (sigma_to**2 * (sigma_from**2 - sigma_to**2) / sigma_from**2) ** 0.5
sigma_down = (sigma_to**2 - sigma_up**2) ** 0.5
# 2. Convert to an ODE derivative
# derivative = (sample - pred_original_sample) / sigma
derivative = model_output
dt = sigma_down - sigma
prev_sample = sample + derivative * dt
device = model_output.device
# noise = randn_tensor(model_output.shape, dtype=model_output.dtype, device=device, generator=generator)
# prev_sample = prev_sample + noise * sigma_up
prev_sample = prev_sample + self.noise_list[self.step_index] * sigma_up
# Cast sample back to model compatible dtype
prev_sample = prev_sample.to(model_output.dtype)
# upon completion increase step index by one
self._step_index += 1
if not return_dict:
return (prev_sample,)
return EulerAncestralDiscreteSchedulerOutput(
prev_sample=prev_sample, pred_original_sample=pred_original_sample
)
def step_and_update_noise(
self,
model_output: torch.FloatTensor,
timestep: Union[float, torch.FloatTensor],
sample: torch.FloatTensor,
expected_prev_sample: torch.FloatTensor,
update_epsilon_type=Epsilon_Update_Type.OVERRIDE,
generator: Optional[torch.Generator] = None,
return_dict: bool = True,
) -> Union[EulerAncestralDiscreteSchedulerOutput, Tuple]:
"""
Predict the sample from the previous timestep by reversing the SDE. This function propagates the diffusion
process from the learned model outputs (most often the predicted noise).
Args:
model_output (`torch.FloatTensor`):
The direct output from learned diffusion model.
timestep (`float`):
The current discrete timestep in the diffusion chain.
sample (`torch.FloatTensor`):
A current instance of a sample created by the diffusion process.
generator (`torch.Generator`, *optional*):
A random number generator.
return_dict (`bool`):
Whether or not to return a
[`~schedulers.scheduling_euler_ancestral_discrete.EulerAncestralDiscreteSchedulerOutput`] or tuple.
Returns:
[`~schedulers.scheduling_euler_ancestral_discrete.EulerAncestralDiscreteSchedulerOutput`] or `tuple`:
If return_dict is `True`,
[`~schedulers.scheduling_euler_ancestral_discrete.EulerAncestralDiscreteSchedulerOutput`] is returned,
otherwise a tuple is returned where the first element is the sample tensor.
"""
if (
isinstance(timestep, int)
or isinstance(timestep, torch.IntTensor)
or isinstance(timestep, torch.LongTensor)
):
raise ValueError(
(
"Passing integer indices (e.g. from `enumerate(timesteps)`) as timesteps to"
" `EulerDiscreteScheduler.step()` is not supported. Make sure to pass"
" one of the `scheduler.timesteps` as a timestep."
),
)
if not self.is_scale_input_called:
logger.warning(
"The `scale_model_input` function should be called before `step` to ensure correct denoising. "
"See `StableDiffusionPipeline` for a usage example."
)
self._init_step_index(timestep.view((1)))
sigma = self.sigmas[self.step_index]
# Upcast to avoid precision issues when computing prev_sample
sample = sample.to(torch.float32)
# 1. compute predicted original sample (x_0) from sigma-scaled predicted noise
if self.config.prediction_type == "epsilon":
pred_original_sample = sample - sigma * model_output
elif self.config.prediction_type == "v_prediction":
# * c_out + input * c_skip
pred_original_sample = model_output * (-sigma / (sigma**2 + 1) ** 0.5) + (sample / (sigma**2 + 1))
elif self.config.prediction_type == "sample":
raise NotImplementedError("prediction_type not implemented yet: sample")
else:
raise ValueError(
f"prediction_type given as {self.config.prediction_type} must be one of `epsilon`, or `v_prediction`"
)
sigma_from = self.sigmas[self.step_index]
sigma_to = self.sigmas[self.step_index + 1]
sigma_up = (sigma_to**2 * (sigma_from**2 - sigma_to**2) / sigma_from**2) ** 0.5
sigma_down = (sigma_to**2 - sigma_up**2) ** 0.5
# 2. Convert to an ODE derivative
# derivative = (sample - pred_original_sample) / sigma
derivative = model_output
dt = sigma_down - sigma
prev_sample = sample + derivative * dt
device = model_output.device
# noise = randn_tensor(model_output.shape, dtype=model_output.dtype, device=device, generator=generator)
# prev_sample = prev_sample + noise * sigma_up
if sigma_up > 0:
req_noise = (expected_prev_sample - prev_sample) / sigma_up
if update_epsilon_type == Epsilon_Update_Type.OVERRIDE:
self.noise_list[self.step_index] = req_noise
else:
for i in range(10):
n = torch.autograd.Variable(self.noise_list[self.step_index].detach().clone(), requires_grad=True)
loss = torch.norm(n - req_noise.detach())
loss.backward()
self.noise_list[self.step_index] -= n.grad.detach() * 1.8
prev_sample = prev_sample + self.noise_list[self.step_index] * sigma_up
# Cast sample back to model compatible dtype
prev_sample = prev_sample.to(model_output.dtype)
# upon completion increase step index by one
self._step_index += 1
if not return_dict:
return (prev_sample,)
return EulerAncestralDiscreteSchedulerOutput(
prev_sample=prev_sample, pred_original_sample=pred_original_sample
)
def inv_step(
self,
model_output: torch.FloatTensor,
timestep: Union[float, torch.FloatTensor],
sample: torch.FloatTensor,
generator: Optional[torch.Generator] = None,
return_dict: bool = True,
) -> Union[EulerAncestralDiscreteSchedulerOutput, Tuple]:
"""
Predict the sample from the previous timestep by reversing the SDE. This function propagates the diffusion
process from the learned model outputs (most often the predicted noise).
Args:
model_output (`torch.FloatTensor`):
The direct output from learned diffusion model.
timestep (`float`):
The current discrete timestep in the diffusion chain.
sample (`torch.FloatTensor`):
A current instance of a sample created by the diffusion process.
generator (`torch.Generator`, *optional*):
A random number generator.
return_dict (`bool`):
Whether or not to return a
[`~schedulers.scheduling_euler_ancestral_discrete.EulerAncestralDiscreteSchedulerOutput`] or tuple.
Returns:
[`~schedulers.scheduling_euler_ancestral_discrete.EulerAncestralDiscreteSchedulerOutput`] or `tuple`:
If return_dict is `True`,
[`~schedulers.scheduling_euler_ancestral_discrete.EulerAncestralDiscreteSchedulerOutput`] is returned,
otherwise a tuple is returned where the first element is the sample tensor.
"""
if (
isinstance(timestep, int)
or isinstance(timestep, torch.IntTensor)
or isinstance(timestep, torch.LongTensor)
):
raise ValueError(
(
"Passing integer indices (e.g. from `enumerate(timesteps)`) as timesteps to"
" `EulerDiscreteScheduler.step()` is not supported. Make sure to pass"
" one of the `scheduler.timesteps` as a timestep."
),
)
if not self.is_scale_input_called:
logger.warning(
"The `scale_model_input` function should be called before `step` to ensure correct denoising. "
"See `StableDiffusionPipeline` for a usage example."
)
self._init_step_index(timestep.view((1)))
sigma = self.sigmas[self.step_index]
# Upcast to avoid precision issues when computing prev_sample
sample = sample.to(torch.float32)
# 1. compute predicted original sample (x_0) from sigma-scaled predicted noise
if self.config.prediction_type == "epsilon":
pred_original_sample = sample - sigma * model_output
elif self.config.prediction_type == "v_prediction":
# * c_out + input * c_skip
pred_original_sample = model_output * (-sigma / (sigma**2 + 1) ** 0.5) + (sample / (sigma**2 + 1))
elif self.config.prediction_type == "sample":
raise NotImplementedError("prediction_type not implemented yet: sample")
else:
raise ValueError(
f"prediction_type given as {self.config.prediction_type} must be one of `epsilon`, or `v_prediction`"
)
sigma_from = self.sigmas[self.step_index]
sigma_to = self.sigmas[self.step_index+1]
# sigma_up = (sigma_to**2 * (sigma_from**2 - sigma_to**2) / sigma_from**2) ** 0.5
sigma_up = (sigma_to**2 * (sigma_from**2 - sigma_to**2).abs() / sigma_from**2) ** 0.5
# sigma_down = (sigma_to**2 - sigma_up**2) ** 0.5
sigma_down = sigma_to**2 / sigma_from
# 2. Convert to an ODE derivative
# derivative = (sample - pred_original_sample) / sigma
derivative = model_output
dt = sigma_down - sigma
# dt = sigma_down - sigma_from
prev_sample = sample - derivative * dt
device = model_output.device
# noise = randn_tensor(model_output.shape, dtype=model_output.dtype, device=device, generator=generator)
# prev_sample = prev_sample + noise * sigma_up
prev_sample = prev_sample - self.noise_list[self.step_index] * sigma_up
# Cast sample back to model compatible dtype
prev_sample = prev_sample.to(model_output.dtype)
# upon completion increase step index by one
self._step_index += 1
if not return_dict:
return (prev_sample,)
return EulerAncestralDiscreteSchedulerOutput(
prev_sample=prev_sample, pred_original_sample=pred_original_sample
)
def get_all_sigmas(self) -> torch.FloatTensor:
sigmas = np.array(((1 - self.alphas_cumprod) / self.alphas_cumprod) ** 0.5)
sigmas = np.concatenate([sigmas[::-1], [0.0]]).astype(np.float32)
return torch.from_numpy(sigmas)
def add_noise_off_schedule(
self,
original_samples: torch.FloatTensor,
noise: torch.FloatTensor,
timesteps: torch.FloatTensor,
) -> torch.FloatTensor:
# Make sure sigmas and timesteps have the same device and dtype as original_samples
sigmas = self.get_all_sigmas()
sigmas = sigmas.to(device=original_samples.device, dtype=original_samples.dtype)
if original_samples.device.type == "mps" and torch.is_floating_point(timesteps):
# mps does not support float64
timesteps = timesteps.to(original_samples.device, dtype=torch.float32)
else:
timesteps = timesteps.to(original_samples.device)
step_indices = 1000 - int(timesteps.item())
sigma = sigmas[step_indices].flatten()
while len(sigma.shape) < len(original_samples.shape):
sigma = sigma.unsqueeze(-1)
noisy_samples = original_samples + noise * sigma
return noisy_samples |