FillandClaude Opus 4.5
37aa6cf9ce
- 4x video upscaling with temporal coherence using Stream-DiffVSR - Model Loader node with precision/device options and xformers support - Upscaler node with chunked processing for memory efficiency - Automatic model download from HuggingFace - Text-guided upscaling support 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
299 lines
12 KiB
Python
299 lines
12 KiB
Python
# Copyright 2024 Stanford University Team and The HuggingFace Team. All rights reserved.
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
|
|
import math
|
|
from dataclasses import dataclass
|
|
from typing import List, Optional, Tuple, Union
|
|
|
|
import numpy as np
|
|
import torch
|
|
|
|
from diffusers.configuration_utils import ConfigMixin, register_to_config
|
|
from diffusers.utils import BaseOutput
|
|
from diffusers.utils.torch_utils import randn_tensor
|
|
from diffusers.schedulers import KarrasDiffusionSchedulers, SchedulerMixin
|
|
|
|
|
|
@dataclass
|
|
class DDIMSchedulerOutput(BaseOutput):
|
|
"""
|
|
Output class for the scheduler's `step` function output.
|
|
|
|
Args:
|
|
prev_sample (`torch.Tensor` 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.Tensor` 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.Tensor
|
|
pred_original_sample: Optional[torch.Tensor] = None
|
|
|
|
|
|
def betas_for_alpha_bar(
|
|
num_diffusion_timesteps,
|
|
max_beta=0.999,
|
|
alpha_transform_type="cosine",
|
|
):
|
|
"""
|
|
Create a beta schedule that discretizes the given alpha_t_bar function.
|
|
"""
|
|
if alpha_transform_type == "cosine":
|
|
def alpha_bar_fn(t):
|
|
return math.cos((t + 0.008) / 1.008 * math.pi / 2) ** 2
|
|
elif alpha_transform_type == "exp":
|
|
def alpha_bar_fn(t):
|
|
return math.exp(t * -12.0)
|
|
else:
|
|
raise ValueError(f"Unsupported alpha_transform_type: {alpha_transform_type}")
|
|
|
|
betas = []
|
|
for i in range(num_diffusion_timesteps):
|
|
t1 = i / num_diffusion_timesteps
|
|
t2 = (i + 1) / num_diffusion_timesteps
|
|
betas.append(min(1 - alpha_bar_fn(t2) / alpha_bar_fn(t1), max_beta))
|
|
return torch.tensor(betas, dtype=torch.float32)
|
|
|
|
|
|
def rescale_zero_terminal_snr(betas):
|
|
"""Rescales betas to have zero terminal SNR."""
|
|
alphas = 1.0 - betas
|
|
alphas_cumprod = torch.cumprod(alphas, dim=0)
|
|
alphas_bar_sqrt = alphas_cumprod.sqrt()
|
|
|
|
alphas_bar_sqrt_0 = alphas_bar_sqrt[0].clone()
|
|
alphas_bar_sqrt_T = alphas_bar_sqrt[-1].clone()
|
|
|
|
alphas_bar_sqrt -= alphas_bar_sqrt_T
|
|
alphas_bar_sqrt *= alphas_bar_sqrt_0 / (alphas_bar_sqrt_0 - alphas_bar_sqrt_T)
|
|
|
|
alphas_bar = alphas_bar_sqrt**2
|
|
alphas = alphas_bar[1:] / alphas_bar[:-1]
|
|
alphas = torch.cat([alphas_bar[0:1], alphas])
|
|
betas = 1 - alphas
|
|
|
|
return betas
|
|
|
|
|
|
class DDIMScheduler(SchedulerMixin, ConfigMixin):
|
|
"""
|
|
DDIMScheduler extends the denoising procedure introduced in DDPMs with non-Markovian guidance.
|
|
"""
|
|
|
|
_compatibles = [e.name for e in KarrasDiffusionSchedulers]
|
|
order = 1
|
|
|
|
@register_to_config
|
|
def __init__(
|
|
self,
|
|
num_train_timesteps: int = 1000,
|
|
beta_start: float = 0.0001,
|
|
beta_end: float = 0.02,
|
|
beta_schedule: str = "linear",
|
|
trained_betas: Optional[Union[np.ndarray, List[float]]] = None,
|
|
clip_sample: bool = True,
|
|
set_alpha_to_one: bool = True,
|
|
steps_offset: int = 0,
|
|
prediction_type: str = "epsilon",
|
|
thresholding: bool = False,
|
|
dynamic_thresholding_ratio: float = 0.995,
|
|
clip_sample_range: float = 1.0,
|
|
sample_max_value: float = 1.0,
|
|
timestep_spacing: str = "leading",
|
|
rescale_betas_zero_snr: bool = False,
|
|
):
|
|
if trained_betas is not None:
|
|
self.betas = torch.tensor(trained_betas, dtype=torch.float32)
|
|
elif beta_schedule == "linear":
|
|
self.betas = torch.linspace(beta_start, beta_end, num_train_timesteps, dtype=torch.float32)
|
|
elif beta_schedule == "scaled_linear":
|
|
self.betas = torch.linspace(beta_start**0.5, beta_end**0.5, num_train_timesteps, dtype=torch.float32) ** 2
|
|
elif beta_schedule == "squaredcos_cap_v2":
|
|
self.betas = betas_for_alpha_bar(num_train_timesteps)
|
|
else:
|
|
raise NotImplementedError(f"{beta_schedule} is not implemented for {self.__class__}")
|
|
|
|
if rescale_betas_zero_snr:
|
|
self.betas = rescale_zero_terminal_snr(self.betas)
|
|
|
|
self.alphas = 1.0 - self.betas
|
|
self.alphas_cumprod = torch.cumprod(self.alphas, dim=0)
|
|
self.final_alpha_cumprod = torch.tensor(1.0) if set_alpha_to_one else self.alphas_cumprod[0]
|
|
self.init_noise_sigma = 1.0
|
|
self.num_inference_steps = None
|
|
self.timesteps = torch.from_numpy(np.arange(0, 1000)[::-1].copy().astype(np.int64))
|
|
self.timestep_spacing = timestep_spacing
|
|
|
|
def scale_model_input(self, sample: torch.Tensor, timestep: Optional[int] = None) -> torch.Tensor:
|
|
return sample
|
|
|
|
def _get_variance(self, timestep, prev_timestep):
|
|
alpha_prod_t = self.alphas_cumprod[timestep]
|
|
alpha_prod_t_prev = self.alphas_cumprod[prev_timestep] if prev_timestep >= 0 else self.final_alpha_cumprod
|
|
beta_prod_t = 1 - alpha_prod_t
|
|
beta_prod_t_prev = 1 - alpha_prod_t_prev
|
|
variance = (beta_prod_t_prev / beta_prod_t) * (1 - alpha_prod_t / alpha_prod_t_prev)
|
|
return variance
|
|
|
|
def _threshold_sample(self, sample: torch.Tensor) -> torch.Tensor:
|
|
dtype = sample.dtype
|
|
batch_size, channels, *remaining_dims = sample.shape
|
|
|
|
if dtype not in (torch.float32, torch.float64):
|
|
sample = sample.float()
|
|
|
|
sample = sample.reshape(batch_size, channels * np.prod(remaining_dims))
|
|
abs_sample = sample.abs()
|
|
s = torch.quantile(abs_sample, self.config.dynamic_thresholding_ratio, dim=1)
|
|
s = torch.clamp(s, min=1, max=self.config.sample_max_value)
|
|
s = s.unsqueeze(1)
|
|
sample = torch.clamp(sample, -s, s) / s
|
|
sample = sample.reshape(batch_size, channels, *remaining_dims)
|
|
sample = sample.to(dtype)
|
|
return sample
|
|
|
|
def set_timesteps(self, num_inference_steps: int = None, device: Union[str, torch.device] = None, timesteps: List[float] = None):
|
|
if timesteps is not None:
|
|
self.timesteps = torch.tensor(timesteps, device=device)
|
|
self.num_inference_steps = len(timesteps)
|
|
return
|
|
|
|
if num_inference_steps > self.config.num_train_timesteps:
|
|
raise ValueError(
|
|
f"`num_inference_steps`: {num_inference_steps} cannot be larger than `self.config.train_timesteps`:"
|
|
f" {self.config.num_train_timesteps}"
|
|
)
|
|
|
|
self.num_inference_steps = num_inference_steps
|
|
|
|
if self.timestep_spacing == "linspace":
|
|
timesteps = (
|
|
np.linspace(0, self.config.num_train_timesteps - 1, num_inference_steps)
|
|
.round()[::-1]
|
|
.copy()
|
|
.astype(np.int64)
|
|
)
|
|
elif self.timestep_spacing == "leading":
|
|
step_ratio = self.config.num_train_timesteps // self.num_inference_steps
|
|
timesteps = (np.arange(0, num_inference_steps) * step_ratio).round()[::-1].copy().astype(np.int64)
|
|
timesteps += self.config.steps_offset
|
|
elif self.timestep_spacing == "trailing":
|
|
step_ratio = self.config.num_train_timesteps / self.num_inference_steps
|
|
timesteps = np.round(np.arange(self.config.num_train_timesteps, 0, -step_ratio)).astype(np.int64)
|
|
timesteps -= 1
|
|
else:
|
|
raise ValueError(
|
|
f"{self.timestep_spacing} is not supported. Please make sure to choose one of 'leading' or 'trailing'."
|
|
)
|
|
|
|
self.timesteps = torch.from_numpy(timesteps).to(device)
|
|
|
|
def step(
|
|
self,
|
|
model_output: torch.Tensor,
|
|
timestep: int,
|
|
sample: torch.Tensor,
|
|
eta: float = 0.0,
|
|
use_clipped_model_output: bool = False,
|
|
generator=None,
|
|
variance_noise: Optional[torch.Tensor] = None,
|
|
return_dict: bool = True,
|
|
) -> Union[DDIMSchedulerOutput, Tuple]:
|
|
if self.num_inference_steps is None:
|
|
raise ValueError(
|
|
"Number of inference steps is 'None', you need to run 'set_timesteps' after creating the scheduler"
|
|
)
|
|
|
|
prev_timestep = timestep - self.config.num_train_timesteps // self.num_inference_steps
|
|
alpha_prod_t = self.alphas_cumprod[timestep]
|
|
alpha_prod_t_prev = self.alphas_cumprod[prev_timestep] if prev_timestep >= 0 else self.final_alpha_cumprod
|
|
beta_prod_t = 1 - alpha_prod_t
|
|
|
|
if self.config.prediction_type == "epsilon":
|
|
pred_original_sample = (sample - beta_prod_t ** (0.5) * model_output) / alpha_prod_t ** (0.5)
|
|
pred_epsilon = model_output
|
|
elif self.config.prediction_type == "sample":
|
|
pred_original_sample = model_output
|
|
pred_epsilon = (sample - alpha_prod_t ** (0.5) * pred_original_sample) / beta_prod_t ** (0.5)
|
|
elif self.config.prediction_type == "v_prediction":
|
|
pred_original_sample = (alpha_prod_t**0.5) * sample - (beta_prod_t**0.5) * model_output
|
|
pred_epsilon = (alpha_prod_t**0.5) * model_output + (beta_prod_t**0.5) * sample
|
|
else:
|
|
raise ValueError(
|
|
f"prediction_type given as {self.config.prediction_type} must be one of `epsilon`, `sample`, or"
|
|
" `v_prediction`"
|
|
)
|
|
|
|
if self.config.thresholding:
|
|
pred_original_sample = self._threshold_sample(pred_original_sample)
|
|
elif self.config.clip_sample:
|
|
pred_original_sample = pred_original_sample.clamp(
|
|
-self.config.clip_sample_range, self.config.clip_sample_range
|
|
)
|
|
|
|
variance = self._get_variance(timestep, prev_timestep)
|
|
std_dev_t = eta * variance ** (0.5)
|
|
|
|
if use_clipped_model_output:
|
|
pred_epsilon = (sample - alpha_prod_t ** (0.5) * pred_original_sample) / beta_prod_t ** (0.5)
|
|
|
|
pred_sample_direction = (1 - alpha_prod_t_prev - std_dev_t**2) ** (0.5) * pred_epsilon
|
|
prev_sample = alpha_prod_t_prev ** (0.5) * pred_original_sample + pred_sample_direction
|
|
|
|
if eta > 0:
|
|
if variance_noise is not None and generator is not None:
|
|
raise ValueError(
|
|
"Cannot pass both generator and variance_noise."
|
|
)
|
|
|
|
if variance_noise is None:
|
|
variance_noise = randn_tensor(
|
|
model_output.shape, generator=generator, device=model_output.device, dtype=model_output.dtype
|
|
)
|
|
variance = std_dev_t * variance_noise
|
|
prev_sample = prev_sample + variance
|
|
|
|
if not return_dict:
|
|
return (prev_sample,)
|
|
|
|
return DDIMSchedulerOutput(prev_sample=prev_sample, pred_original_sample=pred_original_sample)
|
|
|
|
def add_noise(
|
|
self,
|
|
original_samples: torch.Tensor,
|
|
noise: torch.Tensor,
|
|
timesteps: torch.IntTensor,
|
|
) -> torch.Tensor:
|
|
self.alphas_cumprod = self.alphas_cumprod.to(device=original_samples.device)
|
|
alphas_cumprod = self.alphas_cumprod.to(dtype=original_samples.dtype)
|
|
timesteps = timesteps.to(original_samples.device)
|
|
|
|
sqrt_alpha_prod = alphas_cumprod[timesteps] ** 0.5
|
|
sqrt_alpha_prod = sqrt_alpha_prod.flatten()
|
|
while len(sqrt_alpha_prod.shape) < len(original_samples.shape):
|
|
sqrt_alpha_prod = sqrt_alpha_prod.unsqueeze(-1)
|
|
|
|
sqrt_one_minus_alpha_prod = (1 - alphas_cumprod[timesteps]) ** 0.5
|
|
sqrt_one_minus_alpha_prod = sqrt_one_minus_alpha_prod.flatten()
|
|
while len(sqrt_one_minus_alpha_prod.shape) < len(original_samples.shape):
|
|
sqrt_one_minus_alpha_prod = sqrt_one_minus_alpha_prod.unsqueeze(-1)
|
|
|
|
noisy_samples = sqrt_alpha_prod * original_samples + sqrt_one_minus_alpha_prod * noise
|
|
return noisy_samples
|
|
|
|
def __len__(self):
|
|
return self.config.num_train_timesteps
|