models/pipeline_ddpm_text_encoder.py

# Copyright 2024 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 pdb
from typing import List, Optional, Tuple, Union

import torch

from diffusers.utils.torch_utils import randn_tensor
from diffusers.pipelines.pipeline_utils import DiffusionPipeline, ImagePipelineOutput
from transformers import CLIPTextModel, CLIPTokenizer


class DDPMPipeline(DiffusionPipeline):
    r"""
    Pipeline for image generation.

    This model inherits from [`DiffusionPipeline`]. Check the superclass documentation for the generic methods
    implemented for all pipelines (downloading, saving, running on a particular device, etc.).

    Parameters:
        unet ([`UNet2DModel`]):
            A `UNet2DModel` to denoise the encoded image latents.
        scheduler ([`SchedulerMixin`]):
            A scheduler to be used in combination with `unet` to denoise the encoded image. Can be one of
            [`DDPMScheduler`], or [`DDIMScheduler`].
    """

    model_cpu_offload_seq = "unet"

    def __init__(
            self,
            unet,
            scheduler,
            text_encoder: Optional[CLIPTextModel]=None,
            tokenizer: Optional[CLIPTokenizer]=None,
    ):
        super().__init__()

        self.register_modules(
            unet=unet,
            scheduler=scheduler,
            text_encoder=text_encoder,
            tokenizer=tokenizer
        )

    @torch.no_grad()
    def __call__(
        self,
        prompt: Union[str, List[str]] = None,
        generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
        num_inference_steps: int = 1000,
        output_type: Optional[str] = "pil",
        return_dict: bool = True,
        height: Optional[int] = None, # <--- 新增参数
        width: Optional[int] = None,
    ) -> Union[ImagePipelineOutput, Tuple]:
        r"""
        The call function to the pipeline for generation.

        Args:
            batch_size (`int`, *optional*, defaults to 1):
                The number of images to generate.
            generator (`torch.Generator`, *optional*):
                A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
                generation deterministic.
            num_inference_steps (`int`, *optional*, defaults to 1000):
                The number of denoising steps. More denoising steps usually lead to a higher quality image at the
                expense of slower inference.
            output_type (`str`, *optional*, defaults to `"pil"`):
                The output format of the generated image. Choose between `PIL.Image` or `np.array`.
            return_dict (`bool`, *optional*, defaults to `True`):
                Whether or not to return a [`~pipelines.ImagePipelineOutput`] instead of a plain tuple.

        Example:

        ```py
        >>> from diffusers import DDPMPipeline

        >>> # load model and scheduler
        >>> pipe = DDPMPipeline.from_pretrained("google/ddpm-cat-256")

        >>> # run pipeline in inference (sample random noise and denoise)
        >>> image = pipe().images[0]

        >>> # save image
        >>> image.save("ddpm_generated_image.png")
        ```

        Returns:
            [`~pipelines.ImagePipelineOutput`] or `tuple`:
                If `return_dict` is `True`, [`~pipelines.ImagePipelineOutput`] is returned, otherwise a `tuple` is
                returned where the first element is a list with the generated images
        """
        # Prepare prompt embedding
        if prompt is not None and isinstance(prompt, str):
            batch_size = 1
        elif prompt is not None and isinstance(prompt, list):
            batch_size = len(prompt)

        text_inputs = self.tokenizer(
            prompt.lower(),
            padding="max_length",
            max_length=self.tokenizer.model_max_length,
            truncation=True,
            return_tensors="pt",
        )
        text_input_ids = text_inputs.input_ids.to(self.device)
        encoder_hidden_states = self.text_encoder(text_input_ids, return_dict=False)[0]

        if height is None:
            height = self.unet.config.sample_size
        if width is None:
            width = self.unet.config.sample_size
        image_shape = (
            batch_size,
            self.unet.config.in_channels,
            height,
            width,
        )
        # Sample gaussian noise to begin loop
        # if isinstance(self.unet.config.sample_size, int):
        #     image_shape = (
        #         batch_size,
        #         self.unet.config.in_channels,
        #         self.unet.config.sample_size,
        #         self.unet.config.sample_size,
        #     )
        # else:
        #     image_shape = (batch_size, self.unet.config.in_channels, *self.unet.config.sample_size)

        if self.device.type == "mps":
            # randn does not work reproducibly on mps
            image = randn_tensor(image_shape, generator=generator)
            image = image.to(self.device)
        else:
            image = randn_tensor(image_shape, generator=generator, device=self.device)

        # set step values
        self.scheduler.set_timesteps(num_inference_steps)
        for t in self.progress_bar(self.scheduler.timesteps):
            # 1. predict noise model_output
            model_output = self.unet(image, t, encoder_hidden_states).sample
            # 2. compute previous image: x_t -> x_t-1
            image = self.scheduler.step(model_output, t, image, generator=generator).prev_sample
        
        image = (image / 2 + 0.5).clamp(0, 1)
        image = image.cpu().permute(0, 2, 3, 1).numpy()
        if output_type == "pil":
            image = self.numpy_to_pil(image)

        if not return_dict:
            return (image,)

        return ImagePipelineOutput(images=image)