Buckets:
| # AutoPipeline | |
| [AutoPipeline](../api/pipelines/auto_pipeline) 是一种*任务与模型*流水线，它会根据任务自动选择正确的流水线子类。它负责处理加载不同流水线子类的复杂性，无需知道具体的流水线子类名称。 | |
| 这与 [DiffusionPipeline](/docs/diffusers/v0.43.0/zh/api/pipelines/overview#diffusers.DiffusionPipeline) 不同，后者是一个*仅包含模型*的流水线，会根据模型自动选择流水线子类。 | |
| ```text | |
| AutoPipelineForImage2Image.from_pretrained(model_id) | |
| | | |
| +-- read model_index.json (e.g. StableDiffusionXLPipeline) | |
| +-- task mapping (image-to-image) | |
| | | |
| v | |
| StableDiffusionXLImg2ImgPipeline // returned instance | |
| ``` | |
| [AutoPipelineForImage2Image](/docs/diffusers/v0.43.0/zh/api/pipelines/auto_pipeline#diffusers.AutoPipelineForImage2Image) 会返回任务特定的子类（例如 [StableDiffusionXLImg2ImgPipeline](/docs/diffusers/v0.43.0/zh/api/pipelines/stable_diffusion/stable_diffusion_xl#diffusers.StableDiffusionXLImg2ImgPipeline)），该子类仅可用于图像到图像任务。 | |
| ```py | |
| import torch | |
| from diffusers import AutoPipelineForImage2Image | |
| pipeline = AutoPipelineForImage2Image.from_pretrained( | |
| "RunDiffusion/Juggernaut-XL-v9", dtype=torch.bfloat16, device_map="cuda", # or "mps", "xpu", "cpu" | |
| ) | |
| print(pipeline) | |
| # StableDiffusionXLImg2ImgPipeline { | |
| # "_class_name": "StableDiffusionXLImg2ImgPipeline", | |
| # ... | |
| # } | |
| ``` | |
| 使用 [DiffusionPipeline](/docs/diffusers/v0.43.0/zh/api/pipelines/overview#diffusers.DiffusionPipeline) 加载同一模型时，会返回默认的文本到图像子类 [StableDiffusionXLPipeline](/docs/diffusers/v0.43.0/zh/api/pipelines/stable_diffusion/stable_diffusion_xl#diffusers.StableDiffusionXLPipeline)。该流水线用于文本到图像任务。对于图像到图像或修复任务，请加载相应的任务 AutoPipeline，例如 [AutoPipelineForImage2Image](/docs/diffusers/v0.43.0/zh/api/pipelines/auto_pipeline#diffusers.AutoPipelineForImage2Image) 或 [AutoPipelineForInpainting](/docs/diffusers/v0.43.0/zh/api/pipelines/auto_pipeline#diffusers.AutoPipelineForInpainting)，或者对应的特定任务子类。 | |
| ```py | |
| import torch | |
| from diffusers import DiffusionPipeline | |
| pipeline = DiffusionPipeline.from_pretrained( | |
| "RunDiffusion/Juggernaut-XL-v9", dtype=torch.bfloat16, device_map="cuda", # or "mps", "xpu", "cpu" | |
| ) | |
| print(pipeline) | |
| # StableDiffusionXLPipeline { | |
| # "_class_name": "StableDiffusionXLPipeline", | |
| # ... | |
| # } | |
| ``` | |
| ## 使用 from_pipe 切换任务 | |
| 只需加载一次任务 AutoPipeline，随后即可通过 [from_pipe()](/docs/diffusers/v0.43.0/zh/api/pipelines/auto_pipeline#diffusers.AutoPipelineForImage2Image.from_pipe) 切换任务，而无需重新下载权重。组件将从源流水线中复用。 | |
| ```py | |
| import torch | |
| from diffusers import AutoPipelineForText2Image, AutoPipelineForImage2Image | |
| pipeline_t2i = AutoPipelineForText2Image.from_pretrained( | |
| "RunDiffusion/Juggernaut-XL-v9", dtype=torch.bfloat16, device_map="cuda", # or "mps", "xpu", "cpu" | |
| ) | |
| pipeline_i2i = AutoPipelineForImage2Image.from_pipe(pipeline_t2i) | |
| ``` | |
| 有关更多详细信息，请参阅 [在多个流水线中复用模型](../using-diffusers/loading#reusing-models-in-multiple-pipelines)。 | |
| 查看 [mappings](https://github.com/huggingface/diffusers/blob/main/src/diffusers/pipelines/auto_pipeline.py) 以确认模型是否受支持。尝试加载不受支持的模型会返回错误。 | |
| ```py | |
| import torch | |
| from diffusers import AutoPipelineForImage2Image | |
| pipeline = AutoPipelineForImage2Image.from_pretrained( | |
| "openai/shap-e-img2img", dtype=torch.float16, | |
| ) | |
| "ValueError: AutoPipeline can't find a pipeline linked to ShapEImg2ImgPipeline for None" | |
| ``` | |
| 共有四种 [AutoPipeline](../api/pipelines/auto_pipeline) 类： | |
| - [AutoPipelineForText2Image](/docs/diffusers/v0.43.0/zh/api/pipelines/auto_pipeline#diffusers.AutoPipelineForText2Image) | |
| - [AutoPipelineForImage2Image](/docs/diffusers/v0.43.0/zh/api/pipelines/auto_pipeline#diffusers.AutoPipelineForImage2Image) | |
| - [AutoPipelineForInpainting](/docs/diffusers/v0.43.0/zh/api/pipelines/auto_pipeline#diffusers.AutoPipelineForInpainting) | |
| - [AutoPipelineForText2Audio](/docs/diffusers/v0.43.0/zh/api/pipelines/auto_pipeline#diffusers.AutoPipelineForText2Audio) | |
| 这些类中的每一个都包含一个预定义映射，将 pipeline 链接到其任务特定的子类。 | |
| 调用 [from_pretrained()](/docs/diffusers/v0.43.0/zh/api/pipelines/auto_pipeline#diffusers.AutoPipelineForText2Image.from_pretrained) 时，它会从 `model_index.json` 文件中提取类名，并根据映射关系为任务选择相应的 pipeline 子类。 | |
Xet Storage Details
- Size:
- 4.68 kB
- Xet hash:
- ac0dac52fe5badbaf8cdb0ee7bd9c12c79c187589ec6b0267b0eecda60f079be
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.