from typing import Optional
import numpy as np
import torch
import torch.nn.functional as F # noqa
from diffusers import StableDiffusionInpaintPipeline
from diffusers.utils import load_image
from mmengine import print_log
from PIL import Image
from torch import nn
from diffengine.models.editors.stable_diffusion import StableDiffusion
from diffengine.registry import MODELS
@MODELS.register_module()
[docs]class StableDiffusionInpaint(StableDiffusion):
"""Stable Diffusion Inpaint.
Args:
----
model (str): pretrained model name of stable diffusion.
Defaults to 'runwayml/stable-diffusion-v1-5'.
data_preprocessor (dict, optional): The pre-process config of
:class:`SDInpaintDataPreprocessor`.
"""
def __init__(self,
*args,
model: str = "runwayml/stable-diffusion-inpainting",
data_preprocessor: dict | nn.Module | None = None,
**kwargs) -> None:
if data_preprocessor is None:
data_preprocessor = {"type": "SDInpaintDataPreprocessor"}
super().__init__(
*args,
model=model,
data_preprocessor=data_preprocessor,
**kwargs) # type: ignore[misc]
[docs] def prepare_model(self) -> None:
"""Prepare model for training.
Disable gradient for some models.
"""
# Fix input channels of Unet
in_channels = 9
if self.unet.in_channels != in_channels:
out_channels = self.unet.conv_in.out_channels
self.unet.register_to_config(in_channels=in_channels)
with torch.no_grad():
new_conv_in = nn.Conv2d(
in_channels, out_channels, self.unet.conv_in.kernel_size,
self.unet.conv_in.stride, self.unet.conv_in.padding,
)
new_conv_in.weight.zero_()
new_conv_in.weight[:, :4, :, :].copy_(self.unet.conv_in.weight)
self.unet.conv_in = new_conv_in
if self.gradient_checkpointing:
self.unet.enable_gradient_checkpointing()
if self.finetune_text_encoder:
self.text_encoder.gradient_checkpointing_enable()
self.vae.requires_grad_(requires_grad=False)
print_log("Set VAE untrainable.", "current")
if not self.finetune_text_encoder:
self.text_encoder.requires_grad_(requires_grad=False)
print_log("Set Text Encoder untrainable.", "current")
@torch.no_grad()
[docs] def infer(self,
prompt: list[str],
image: list[str | Image.Image],
mask: list[str | Image.Image],
negative_prompt: str | None = None,
height: int | None = None,
width: int | None = None,
num_inference_steps: int = 50,
output_type: str = "pil",
**kwargs) -> list[np.ndarray]:
"""Inference function.
Args:
----
prompt (`List[str]`):
The prompt or prompts to guide the image generation.
image (`List[Union[str, Image.Image]]`):
The image for inpainting.
mask (`List[Union[str, Image.Image]]`):
The mask for inpainting.
negative_prompt (`Optional[str]`):
The prompt or prompts to guide the image generation.
Defaults to None.
height (int, optional):
The height in pixels of the generated image. Defaults to None.
width (int, optional):
The width in pixels of the generated image. Defaults to None.
num_inference_steps (int): Number of inference steps.
Defaults to 50.
output_type (str): The output format of the generate image.
Choose between 'pil' and 'latent'. Defaults to 'pil'.
**kwargs: Other arguments.
"""
assert len(prompt) == len(image) == len(mask)
pipeline = StableDiffusionInpaintPipeline.from_pretrained(
self.model,
vae=self.vae,
text_encoder=self.text_encoder,
tokenizer=self.tokenizer,
unet=self.unet,
safety_checker=None,
torch_dtype=(torch.float16 if self.device != torch.device("cpu")
else torch.float32),
)
if self.prediction_type is not None:
# set prediction_type of scheduler if defined
scheduler_args = {"prediction_type": self.prediction_type}
pipeline.scheduler = pipeline.scheduler.from_config(
pipeline.scheduler.config, **scheduler_args)
pipeline.set_progress_bar_config(disable=True)
images = []
for p, img, m in zip(prompt, image, mask, strict=True):
pil_img = load_image(img) if isinstance(img, str) else img
pil_img = pil_img.convert("RGB")
mask_image = load_image(m) if isinstance(m, str) else m
mask_image = mask_image.convert("L")
image = pipeline(
p,
mask_image=mask_image,
image=pil_img,
negative_prompt=negative_prompt,
num_inference_steps=num_inference_steps,
height=height,
width=width,
output_type=output_type,
**kwargs).images[0]
if output_type == "latent":
images.append(image)
else:
images.append(np.array(image))
del pipeline
torch.cuda.empty_cache()
return images
[docs] def forward(
self,
inputs: dict,
data_samples: Optional[list] = None, # noqa
mode: str = "loss") -> dict:
"""Forward function.
Args:
----
inputs (dict): The input dict.
data_samples (Optional[list], optional): The data samples.
Defaults to None.
mode (str, optional): The mode. Defaults to "loss".
Returns:
-------
dict: The loss dict.
"""
assert mode == "loss"
inputs["text"] = self.tokenizer(
inputs["text"],
max_length=self.tokenizer.model_max_length,
padding="max_length",
truncation=True,
return_tensors="pt").input_ids.to(self.device)
num_batches = len(inputs["img"])
if "result_class_image" in inputs:
# use prior_loss_weight
weight = torch.cat([
torch.ones((num_batches // 2, )),
torch.ones((num_batches // 2, )) * self.prior_loss_weight,
]).to(self.device).float().reshape(-1, 1, 1, 1)
else:
weight = None
latents = self._forward_vae(inputs["img"], num_batches)
masked_latents = self._forward_vae(inputs["masked_image"], num_batches)
mask = F.interpolate(inputs["mask"],
size=(latents.shape[2], latents.shape[3]))
noise = self.noise_generator(latents)
timesteps = self.timesteps_generator(self.scheduler, num_batches,
self.device)
noisy_latents = self._preprocess_model_input(latents, noise, timesteps)
latent_model_input = torch.cat([noisy_latents, mask, masked_latents], dim=1)
encoder_hidden_states = self.text_encoder(inputs["text"])[0]
model_pred = self.unet(
latent_model_input,
timesteps,
encoder_hidden_states=encoder_hidden_states).sample
return self.loss(model_pred, noise, latents, timesteps, weight)