Diffusers documentation
SequentialPipelineBlocks
SequentialPipelineBlocks
SequentialPipelineBlocks 是一种多块类型，它将其他 ModularPipelineBlocks 按顺序组合在一起。数据通过 inputs 和 intermediate_outputs 从一个块线性流向下一个块。SequentialPipelineBlocks 中的每个块通常代表流水线中的一个步骤，通过将它们组合起来，你可以逐步构建出完整的流水线。
本指南将向您展示如何将两个块连接到一个 SequentialPipelineBlocks 中。
创建两个 ModularPipelineBlocks。第一个块 InputBlock 输出一个 batch_size 值，第二个块 ImageEncoderBlock 将 batch_size 用作 inputs。
from diffusers.modular_pipelines import ModularPipelineBlocks, InputParam, OutputParam
class InputBlock(ModularPipelineBlocks):
@property
def inputs(self):
return [
InputParam(name="prompt", type_hint=list, description="list of text prompts"),
InputParam(name="num_images_per_prompt", type_hint=int, description="number of images per prompt"),
]
@property
def intermediate_outputs(self):
return [
OutputParam(name="batch_size", description="calculated batch size"),
]
@property
def description(self):
return "A block that determines batch_size based on the number of prompts and num_images_per_prompt argument."
def __call__(self, components, state):
block_state = self.get_block_state(state)
batch_size = len(block_state.prompt)
block_state.batch_size = batch_size * block_state.num_images_per_prompt
self.set_block_state(state, block_state)
return components, stateimport torch
from diffusers.modular_pipelines import ModularPipelineBlocks, InputParam, OutputParam
class ImageEncoderBlock(ModularPipelineBlocks):
@property
def inputs(self):
return [
InputParam(name="image", type_hint="PIL.Image", description="raw input image to process"),
InputParam(name="batch_size", type_hint=int),
]
@property
def intermediate_outputs(self):
return [
OutputParam(name="image_latents", description="latents representing the image"),
]
@property
def description(self):
return "Encode raw image into its latent presentation"
def __call__(self, components, state):
block_state = self.get_block_state(state)
# Simulate processing the image
# This will change the state of the image from a PIL image to a tensor for all blocks
block_state.image = torch.randn(1, 3, 512, 512)
block_state.batch_size = block_state.batch_size * 2
block_state.image_latents = torch.randn(1, 4, 64, 64)
self.set_block_state(state, block_state)
return components, state通过定义 SequentialPipelineBlocks 来连接这两个块。在 block_classes 中列出块实例，并在 block_names 中列出它们对应的名称。这些块将按照其在 block_classes 中出现的顺序执行，数据通过 PipelineState 从一个块流向下一个块。
class ImageProcessingStep(SequentialPipelineBlocks):
"""
# auto_docstring
"""
model_name = "my_model"
block_classes = [InputBlock(), ImageEncoderBlock()]
block_names = ["input", "image_encoder"]
@property
def description(self):
return (
"Process text prompts and images for the pipeline. It:\n"
" - Determines the batch size from the prompts.\n"
" - Encodes the image into latent space."
)创建 SequentialPipelineBlocks 时，诸如 inputs、intermediate_outputs 和 expected_components 等属性会自动从子块中聚合，因此无需再次定义。
您应该设置以下几个属性：
description：我们建议为组装后的块添加描述，以说明该组合步骤的作用。model_name：此参数会自动从子块中推导得出，但结果并不总是正确的，因此您可能需要手动覆盖它。outputs：默认情况下，它与intermediate_outputs相同，但您可以手动设置它以控制文档中显示哪些值。这对于仅显示最终输出而非所有中间值非常有用。
这些属性连同聚合后的 inputs、intermediate_outputs 和 expected_components，用于自动生成 doc 属性。
打印 ImageProcessingStep 块以检查其子块，并使用 doc 查看该块的输入、输出和组件的完整摘要。
blocks = ImageProcessingStep()
print(blocks)
print(blocks.doc)