update

src/diffusers/pipelines/amused/pipeline_amused.py

@@ -21,7 +21,7 @@ from ...image_processor import VaeImageProcessor
 from ...models import UVit2DModel, VQModel
 from ...schedulers import AmusedScheduler
 from ...utils import is_torch_xla_available, replace_example_docstring
-from ..pipeline_utils import DiffusionPipeline, ImagePipelineOutput
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, ImagePipelineOutput
 
 
 if is_torch_xla_available():
@@ -47,7 +47,8 @@ EXAMPLE_DOC_STRING = """
 """
 
 
-class AmusedPipeline(DiffusionPipeline):
+class AmusedPipeline(DeprecatedPipelineMixin, DiffusionPipeline):
+    _last_supported_version = "0.33.1"
     image_processor: VaeImageProcessor
     vqvae: VQModel
     tokenizer: CLIPTokenizer

src/diffusers/pipelines/amused/pipeline_amused_img2img.py

@@ -21,7 +21,7 @@ from ...image_processor import PipelineImageInput, VaeImageProcessor
 from ...models import UVit2DModel, VQModel
 from ...schedulers import AmusedScheduler
 from ...utils import is_torch_xla_available, replace_example_docstring
-from ..pipeline_utils import DiffusionPipeline, ImagePipelineOutput
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, ImagePipelineOutput
 
 
 if is_torch_xla_available():
@@ -57,7 +57,8 @@ EXAMPLE_DOC_STRING = """
 """
 
 
-class AmusedImg2ImgPipeline(DiffusionPipeline):
+class AmusedImg2ImgPipeline(DeprecatedPipelineMixin, DiffusionPipeline):
+    _last_supported_version = "0.33.1"
     image_processor: VaeImageProcessor
     vqvae: VQModel
     tokenizer: CLIPTokenizer

src/diffusers/pipelines/amused/pipeline_amused_inpaint.py

@@ -22,7 +22,7 @@ from ...image_processor import PipelineImageInput, VaeImageProcessor
 from ...models import UVit2DModel, VQModel
 from ...schedulers import AmusedScheduler
 from ...utils import is_torch_xla_available, replace_example_docstring
-from ..pipeline_utils import DiffusionPipeline, ImagePipelineOutput
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, ImagePipelineOutput
 
 
 if is_torch_xla_available():
@@ -65,7 +65,8 @@ EXAMPLE_DOC_STRING = """
 """
 
 
-class AmusedInpaintPipeline(DiffusionPipeline):
+class AmusedInpaintPipeline(DeprecatedPipelineMixin, DiffusionPipeline):
+    _last_supported_version = "0.33.1"
     image_processor: VaeImageProcessor
     vqvae: VQModel
     tokenizer: CLIPTokenizer

src/diffusers/pipelines/audioldm/pipeline_audioldm.py

@@ -24,7 +24,7 @@ from ...models import AutoencoderKL, UNet2DConditionModel
 from ...schedulers import KarrasDiffusionSchedulers
 from ...utils import is_torch_xla_available, logging, replace_example_docstring
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import AudioPipelineOutput, DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import AudioPipelineOutput, DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 
 
 if is_torch_xla_available():
@@ -57,7 +57,7 @@ EXAMPLE_DOC_STRING = """
 """
 
 
-class AudioLDMPipeline(DiffusionPipeline, StableDiffusionMixin):
+class AudioLDMPipeline(DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin):
     r"""
     Pipeline for text-to-audio generation using AudioLDM.
 
@@ -81,6 +81,7 @@ class AudioLDMPipeline(DiffusionPipeline, StableDiffusionMixin):
             Vocoder of class `SpeechT5HifiGan`.
     """
 
+    _last_supported_version = "0.33.1"
     model_cpu_offload_seq = "text_encoder->unet->vae"
 
     def __init__(

src/diffusers/pipelines/blip_diffusion/pipeline_blip_diffusion.py

@@ -25,7 +25,7 @@ from ...utils import (
     replace_example_docstring,
 )
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, ImagePipelineOutput
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, ImagePipelineOutput
 from .blip_image_processing import BlipImageProcessor
 from .modeling_blip2 import Blip2QFormerModel
 from .modeling_ctx_clip import ContextCLIPTextModel
@@ -81,7 +81,7 @@ EXAMPLE_DOC_STRING = """
 """
 
 
-class BlipDiffusionPipeline(DiffusionPipeline):
+class BlipDiffusionPipeline(DeprecatedPipelineMixin, DiffusionPipeline):
     """
     Pipeline for Zero-Shot Subject Driven Generation using Blip Diffusion.
 
@@ -107,6 +107,7 @@ class BlipDiffusionPipeline(DiffusionPipeline):
             Position of the context token in the text encoder.
     """
 
+    _last_supported_version = "0.33.1"
     model_cpu_offload_seq = "qformer->text_encoder->unet->vae"
 
     def __init__(

src/diffusers/pipelines/controlnet_xs/pipeline_controlnet_xs.py

@@ -37,7 +37,7 @@ from ...utils import (
     unscale_lora_layers,
 )
 from ...utils.torch_utils import is_compiled_module, is_torch_version, randn_tensor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 from ..stable_diffusion.pipeline_output import StableDiffusionPipelineOutput
 from ..stable_diffusion.safety_checker import StableDiffusionSafetyChecker
 
@@ -98,6 +98,7 @@ EXAMPLE_DOC_STRING = """
 
 
 class StableDiffusionControlNetXSPipeline(
+    DeprecatedPipelineMixin,
     DiffusionPipeline,
     StableDiffusionMixin,
     TextualInversionLoaderMixin,
@@ -138,6 +139,7 @@ class StableDiffusionControlNetXSPipeline(
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
 
+    _last_supported_version = "0.33.1"
     model_cpu_offload_seq = "text_encoder->unet->vae"
     _optional_components = ["safety_checker", "feature_extractor"]
     _exclude_from_cpu_offload = ["safety_checker"]

src/diffusers/pipelines/controlnet_xs/pipeline_controlnet_xs_sd_xl.py

@@ -46,7 +46,7 @@ from ...utils import (
     unscale_lora_layers,
 )
 from ...utils.torch_utils import is_compiled_module, is_torch_version, randn_tensor
-from ..pipeline_utils import DiffusionPipeline
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline
 from ..stable_diffusion_xl.pipeline_output import StableDiffusionXLPipelineOutput
 
 
@@ -114,6 +114,7 @@ EXAMPLE_DOC_STRING = """
 
 
 class StableDiffusionXLControlNetXSPipeline(
+    DeprecatedPipelineMixin,
     DiffusionPipeline,
     TextualInversionLoaderMixin,
     StableDiffusionXLLoraLoaderMixin,
@@ -158,6 +159,7 @@ class StableDiffusionXLControlNetXSPipeline(
             watermarker is used.
     """
 
+    _last_supported_version = "0.33.1"
     model_cpu_offload_seq = "text_encoder->text_encoder_2->unet->vae"
     _optional_components = [
         "tokenizer",

src/diffusers/pipelines/dance_diffusion/pipeline_dance_diffusion.py

@@ -21,7 +21,7 @@ from ...models import UNet1DModel
 from ...schedulers import SchedulerMixin
 from ...utils import is_torch_xla_available, logging
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import AudioPipelineOutput, DiffusionPipeline
+from ..pipeline_utils import AudioPipelineOutput, DeprecatedPipelineMixin, DiffusionPipeline
 
 
 if is_torch_xla_available():
@@ -34,7 +34,7 @@ else:
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 
 
-class DanceDiffusionPipeline(DiffusionPipeline):
+class DanceDiffusionPipeline(DeprecatedPipelineMixin, DiffusionPipeline):
     r"""
     Pipeline for audio generation.
 
@@ -49,6 +49,7 @@ class DanceDiffusionPipeline(DiffusionPipeline):
             [`IPNDMScheduler`].
     """
 
+    _last_supported_version = "0.33.1"
     model_cpu_offload_seq = "unet"
 
     def __init__(self, unet: UNet1DModel, scheduler: SchedulerMixin):

src/diffusers/pipelines/i2vgen_xl/pipeline_i2vgen_xl.py

@@ -33,7 +33,7 @@ from ...utils import (
 )
 from ...utils.torch_utils import randn_tensor
 from ...video_processor import VideoProcessor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 
 
 if is_torch_xla_available():
@@ -97,9 +97,11 @@ class I2VGenXLPipelineOutput(BaseOutput):
 
 
 class I2VGenXLPipeline(
+    DeprecatedPipelineMixin,
     DiffusionPipeline,
     StableDiffusionMixin,
 ):
+    _last_supported_version = "0.33.1"
     r"""
     Pipeline for image-to-video generation as proposed in [I2VGenXL](https://i2vgen-xl.github.io/).
 

src/diffusers/pipelines/musicldm/pipeline_musicldm.py

@@ -36,7 +36,7 @@ from ...utils import (
     replace_example_docstring,
 )
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import AudioPipelineOutput, DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import AudioPipelineOutput, DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 
 
 if is_librosa_available():
@@ -76,7 +76,8 @@ EXAMPLE_DOC_STRING = """
 """
 
 
-class MusicLDMPipeline(DiffusionPipeline, StableDiffusionMixin):
+class MusicLDMPipeline(DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin):
+    _last_supported_version = "0.33.1"
     r"""
     Pipeline for text-to-audio generation using MusicLDM.
 

src/diffusers/pipelines/paint_by_example/pipeline_paint_by_example.py

@@ -25,7 +25,7 @@ from ...models import AutoencoderKL, UNet2DConditionModel
 from ...schedulers import DDIMScheduler, LMSDiscreteScheduler, PNDMScheduler
 from ...utils import deprecate, is_torch_xla_available, logging
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 from ..stable_diffusion import StableDiffusionPipelineOutput
 from ..stable_diffusion.safety_checker import StableDiffusionSafetyChecker
 from .image_encoder import PaintByExampleImageEncoder
@@ -155,7 +155,8 @@ def prepare_mask_and_masked_image(image, mask):
     return mask, masked_image
 
 
-class PaintByExamplePipeline(DiffusionPipeline, StableDiffusionMixin):
+class PaintByExamplePipeline(DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin):
+    _last_supported_version = "0.33.1"
     r"""
     <Tip warning={true}>
 

src/diffusers/pipelines/pia/pipeline_pia.py

@@ -46,7 +46,7 @@ from ...utils import (
 from ...utils.torch_utils import randn_tensor
 from ...video_processor import VideoProcessor
 from ..free_init_utils import FreeInitMixin
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 
 
 if is_torch_xla_available():
@@ -132,6 +132,7 @@ class PIAPipelineOutput(BaseOutput):
 
 
 class PIAPipeline(
+    DeprecatedPipelineMixin,
     DiffusionPipeline,
     StableDiffusionMixin,
     TextualInversionLoaderMixin,
@@ -140,6 +141,7 @@ class PIAPipeline(
     FromSingleFileMixin,
     FreeInitMixin,
 ):
+    _last_supported_version = "0.33.1"
     r"""
     Pipeline for text-to-video generation.
 

src/diffusers/pipelines/pipeline_utils.py

@@ -139,6 +139,43 @@ class AudioPipelineOutput(BaseOutput):
     audios: np.ndarray
 
 
+class DeprecatedPipelineMixin:
+    """
+    A mixin that can be used to mark a pipeline as deprecated.
+
+    Pipelines inheriting from this mixin will raise a warning when instantiated, indicating that they are deprecated
+    and won't receive updates past the specified version. Tests will be skipped for pipelines that inherit from this
+    mixin.
+
+    Example usage:
+    ```python
+    class MyDeprecatedPipeline(DeprecatedPipelineMixin, DiffusionPipeline):
+        _last_supported_version = "0.20.0"
+
+        def __init__(self, *args, **kwargs):
+            super().__init__(*args, **kwargs)
+    ```
+    """
+
+    # Override this in the inheriting class to specify the last version that will support this pipeline
+    _last_supported_version = None
+
+    def __init__(self, *args, **kwargs):
+        # Get the class name for the warning message
+        class_name = self.__class__.__name__
+
+        # Get the last supported version or use the current version if not specified
+        version_info = getattr(self.__class__, "_last_supported_version", __version__)
+
+        # Raise a warning that this pipeline is deprecated
+        logger.warning(
+            f"The {class_name} has been deprecated and will not receive bug fixes or feature updates after Diffusers version {version_info}. "
+        )
+
+        # Call the parent class's __init__ method
+        super().__init__(*args, **kwargs)
+
+
 class DiffusionPipeline(ConfigMixin, PushToHubMixin):
     r"""
     Base class for all pipelines.

src/diffusers/pipelines/semantic_stable_diffusion/pipeline_semantic_stable_diffusion.py

@@ -11,7 +11,7 @@ from ...pipelines.stable_diffusion.safety_checker import StableDiffusionSafetyCh
 from ...schedulers import KarrasDiffusionSchedulers
 from ...utils import deprecate, is_torch_xla_available, logging
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 from .pipeline_output import SemanticStableDiffusionPipelineOutput
 
 
@@ -25,7 +25,8 @@ else:
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 
 
-class SemanticStableDiffusionPipeline(DiffusionPipeline, StableDiffusionMixin):
+class SemanticStableDiffusionPipeline(DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin):
+    _last_supported_version = "0.33.1"
     r"""
     Pipeline for text-to-image generation using Stable Diffusion with latent editing.
 

src/diffusers/pipelines/stable_diffusion_attend_and_excite/pipeline_stable_diffusion_attend_and_excite.py

@@ -37,7 +37,7 @@ from ...utils import (
     unscale_lora_layers,
 )
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 from ..stable_diffusion import StableDiffusionPipelineOutput
 from ..stable_diffusion.safety_checker import StableDiffusionSafetyChecker
 
@@ -179,7 +179,9 @@ class AttendExciteAttnProcessor:
         return hidden_states
 
 
-class StableDiffusionAttendAndExcitePipeline(DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin):
+class StableDiffusionAttendAndExcitePipeline(
+    DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin
+):
     r"""
     Pipeline for text-to-image generation using Stable Diffusion and Attend-and-Excite.
 
@@ -209,6 +211,8 @@ class StableDiffusionAttendAndExcitePipeline(DiffusionPipeline, StableDiffusionM
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
 
+    _last_supported_version = "0.33.1"
+
     model_cpu_offload_seq = "text_encoder->unet->vae"
     _optional_components = ["safety_checker", "feature_extractor"]
     _exclude_from_cpu_offload = ["safety_checker"]

src/diffusers/pipelines/stable_diffusion_diffedit/pipeline_stable_diffusion_diffedit.py

@@ -40,7 +40,7 @@ from ...utils import (
     unscale_lora_layers,
 )
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 from ..stable_diffusion import StableDiffusionPipelineOutput
 from ..stable_diffusion.safety_checker import StableDiffusionSafetyChecker
 
@@ -242,7 +242,11 @@ def preprocess_mask(mask, batch_size: int = 1):
 
 
 class StableDiffusionDiffEditPipeline(
-    DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin, StableDiffusionLoraLoaderMixin
+    DeprecatedPipelineMixin,
+    DiffusionPipeline,
+    StableDiffusionMixin,
+    TextualInversionLoaderMixin,
+    StableDiffusionLoraLoaderMixin,
 ):
     r"""
     <Tip warning={true}>
@@ -282,6 +286,8 @@ class StableDiffusionDiffEditPipeline(
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
 
+    _last_supported_version = "0.33.1"
+
     model_cpu_offload_seq = "text_encoder->unet->vae"
     _optional_components = ["safety_checker", "feature_extractor", "inverse_scheduler"]
     _exclude_from_cpu_offload = ["safety_checker"]

src/diffusers/pipelines/stable_diffusion_gligen/pipeline_stable_diffusion_gligen.py

@@ -36,7 +36,7 @@ from ...utils import (
     unscale_lora_layers,
 )
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 from ..stable_diffusion import StableDiffusionPipelineOutput
 from ..stable_diffusion.safety_checker import StableDiffusionSafetyChecker
 
@@ -108,7 +108,7 @@ EXAMPLE_DOC_STRING = """
 """
 
 
-class StableDiffusionGLIGENPipeline(DiffusionPipeline, StableDiffusionMixin):
+class StableDiffusionGLIGENPipeline(DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin):
     r"""
     Pipeline for text-to-image generation using Stable Diffusion with Grounded-Language-to-Image Generation (GLIGEN).
 
@@ -135,6 +135,8 @@ class StableDiffusionGLIGENPipeline(DiffusionPipeline, StableDiffusionMixin):
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
 
+    _last_supported_version = "0.33.1"
+
     _optional_components = ["safety_checker", "feature_extractor"]
     model_cpu_offload_seq = "text_encoder->unet->vae"
     _exclude_from_cpu_offload = ["safety_checker"]

src/diffusers/pipelines/stable_diffusion_gligen/pipeline_stable_diffusion_gligen_text_image.py

@@ -41,7 +41,7 @@ from ...utils import (
     unscale_lora_layers,
 )
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 from ..stable_diffusion import StableDiffusionPipelineOutput
 from ..stable_diffusion.clip_image_project_model import CLIPImageProjection
 from ..stable_diffusion.safety_checker import StableDiffusionSafetyChecker
@@ -160,7 +160,7 @@ EXAMPLE_DOC_STRING = """
 """
 
 
-class StableDiffusionGLIGENTextImagePipeline(DiffusionPipeline, StableDiffusionMixin):
+class StableDiffusionGLIGENTextImagePipeline(DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin):
     r"""
     Pipeline for text-to-image generation using Stable Diffusion with Grounded-Language-to-Image Generation (GLIGEN).
 
@@ -193,6 +193,8 @@ class StableDiffusionGLIGENTextImagePipeline(DiffusionPipeline, StableDiffusionM
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
 
+    _last_supported_version = "0.33.1"
+
     model_cpu_offload_seq = "text_encoder->unet->vae"
     _optional_components = ["safety_checker", "feature_extractor"]
     _exclude_from_cpu_offload = ["safety_checker"]

src/diffusers/pipelines/stable_diffusion_k_diffusion/pipeline_stable_diffusion_k_diffusion.py

@@ -42,7 +42,7 @@ from ...utils import (
     unscale_lora_layers,
 )
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 from ..stable_diffusion import StableDiffusionPipelineOutput, StableDiffusionSafetyChecker
 
 
@@ -64,7 +64,11 @@ class ModelWrapper:
 
 
 class StableDiffusionKDiffusionPipeline(
-    DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin, StableDiffusionLoraLoaderMixin
+    DeprecatedPipelineMixin,
+    DiffusionPipeline,
+    StableDiffusionMixin,
+    TextualInversionLoaderMixin,
+    StableDiffusionLoraLoaderMixin,
 ):
     r"""
     Pipeline for text-to-image generation using Stable Diffusion.
@@ -105,6 +109,8 @@ class StableDiffusionKDiffusionPipeline(
             Model that extracts features from generated images to be used as inputs for the `safety_checker`.
     """
 
+    _last_supported_version = "0.33.1"
+
     model_cpu_offload_seq = "text_encoder->unet->vae"
     _optional_components = ["safety_checker", "feature_extractor"]
     _exclude_from_cpu_offload = ["safety_checker"]

src/diffusers/pipelines/stable_diffusion_k_diffusion/pipeline_stable_diffusion_xl_k_diffusion.py

@@ -48,7 +48,7 @@ from ...utils import (
     unscale_lora_layers,
 )
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 from ..stable_diffusion_xl.pipeline_output import StableDiffusionXLPipelineOutput
 
 
@@ -88,6 +88,7 @@ class ModelWrapper:
 
 
 class StableDiffusionXLKDiffusionPipeline(
+    DeprecatedPipelineMixin,
     DiffusionPipeline,
     StableDiffusionMixin,
     FromSingleFileMixin,
@@ -95,6 +96,8 @@ class StableDiffusionXLKDiffusionPipeline(
     TextualInversionLoaderMixin,
     IPAdapterMixin,
 ):
+    _last_supported_version = "0.33.1"
+
     r"""
     Pipeline for text-to-image generation using Stable Diffusion XL and k-diffusion.
 

src/diffusers/pipelines/stable_diffusion_ldm3d/pipeline_stable_diffusion_ldm3d.py

@@ -37,7 +37,7 @@ from ...utils import (
     unscale_lora_layers,
 )
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 from ..stable_diffusion.safety_checker import StableDiffusionSafetyChecker
 
 
@@ -178,6 +178,7 @@ class LDM3DPipelineOutput(BaseOutput):
 
 
 class StableDiffusionLDM3DPipeline(
+    DeprecatedPipelineMixin,
     DiffusionPipeline,
     StableDiffusionMixin,
     TextualInversionLoaderMixin,
@@ -185,6 +186,8 @@ class StableDiffusionLDM3DPipeline(
     StableDiffusionLoraLoaderMixin,
     FromSingleFileMixin,
 ):
+    _last_supported_version = "0.33.1"
+
     r"""
     Pipeline for text-to-image and 3D generation using LDM3D.
 

src/diffusers/pipelines/stable_diffusion_panorama/pipeline_stable_diffusion_panorama.py

@@ -33,7 +33,7 @@ from ...utils import (
     unscale_lora_layers,
 )
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 from ..stable_diffusion import StableDiffusionPipelineOutput
 from ..stable_diffusion.safety_checker import StableDiffusionSafetyChecker
 
@@ -156,12 +156,15 @@ def retrieve_timesteps(
 
 
 class StableDiffusionPanoramaPipeline(
+    DeprecatedPipelineMixin,
     DiffusionPipeline,
     StableDiffusionMixin,
     TextualInversionLoaderMixin,
     StableDiffusionLoraLoaderMixin,
     IPAdapterMixin,
 ):
+    _last_supported_version = "0.33.1"
+
     r"""
     Pipeline for text-to-image generation using MultiDiffusion.
 

src/diffusers/pipelines/stable_diffusion_safe/pipeline_stable_diffusion_safe.py

@@ -14,7 +14,7 @@ from ...models import AutoencoderKL, ImageProjection, UNet2DConditionModel
 from ...schedulers import KarrasDiffusionSchedulers
 from ...utils import deprecate, is_torch_xla_available, logging
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 from . import StableDiffusionSafePipelineOutput
 from .safety_checker import SafeStableDiffusionSafetyChecker
 
@@ -29,7 +29,9 @@ else:
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 
 
-class StableDiffusionPipelineSafe(DiffusionPipeline, StableDiffusionMixin, IPAdapterMixin):
+class StableDiffusionPipelineSafe(DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin, IPAdapterMixin):
+    _last_supported_version = "0.33.1"
+
     r"""
     Pipeline based on the [`StableDiffusionPipeline`] for text-to-image generation using Safe Latent Diffusion.
 

src/diffusers/pipelines/stable_diffusion_sag/pipeline_stable_diffusion_sag.py

@@ -34,7 +34,7 @@ from ...utils import (
     unscale_lora_layers,
 )
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 from ..stable_diffusion import StableDiffusionPipelineOutput
 from ..stable_diffusion.safety_checker import StableDiffusionSafetyChecker
 
@@ -107,7 +107,11 @@ class CrossAttnStoreProcessor:
 
 
 # Modified to get self-attention guidance scale in this paper (https://huggingface.co/papers/2210.00939) as an input
-class StableDiffusionSAGPipeline(DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin, IPAdapterMixin):
+class StableDiffusionSAGPipeline(
+    DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin, IPAdapterMixin
+):
+    _last_supported_version = "0.33.1"
+
     r"""
     Pipeline for text-to-image generation using Stable Diffusion.
 

src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth.py

@@ -33,7 +33,7 @@ from ...utils import (
 )
 from ...utils.torch_utils import randn_tensor
 from ...video_processor import VideoProcessor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 from . import TextToVideoSDPipelineOutput
 
 
@@ -68,8 +68,13 @@ EXAMPLE_DOC_STRING = """
 
 
 class TextToVideoSDPipeline(
-    DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin, StableDiffusionLoraLoaderMixin
+    DeprecatedPipelineMixin,
+    DiffusionPipeline,
+    StableDiffusionMixin,
+    TextualInversionLoaderMixin,
+    StableDiffusionLoraLoaderMixin,
 ):
+    _last_supported_version = "0.33.1"
     r"""
     Pipeline for text-to-video generation.
 

src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth_img2img.py

@@ -34,7 +34,7 @@ from ...utils import (
 )
 from ...utils.torch_utils import randn_tensor
 from ...video_processor import VideoProcessor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 from . import TextToVideoSDPipelineOutput
 
 
@@ -103,8 +103,13 @@ def retrieve_latents(
 
 
 class VideoToVideoSDPipeline(
-    DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin, StableDiffusionLoraLoaderMixin
+    DeprecatedPipelineMixin,
+    DiffusionPipeline,
+    StableDiffusionMixin,
+    TextualInversionLoaderMixin,
+    StableDiffusionLoraLoaderMixin,
 ):
+    _last_supported_version = "0.33.1"
     r"""
     Pipeline for text-guided video-to-video generation.
 

src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_zero.py

@@ -24,7 +24,7 @@ from ...utils import (
     unscale_lora_layers,
 )
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 from ..stable_diffusion import StableDiffusionSafetyChecker
 
 
@@ -296,12 +296,14 @@ def create_motion_field_and_warp_latents(motion_field_strength_x, motion_field_s
 
 
 class TextToVideoZeroPipeline(
+    DeprecatedPipelineMixin,
     DiffusionPipeline,
     StableDiffusionMixin,
     TextualInversionLoaderMixin,
     StableDiffusionLoraLoaderMixin,
     FromSingleFileMixin,
 ):
+    _last_supported_version = "0.33.1"
     r"""
     Pipeline for zero-shot text-to-video generation using Stable Diffusion.
 

src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_zero_sdxl.py

@@ -35,7 +35,7 @@ from ...utils import (
     unscale_lora_layers,
 )
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin
 
 
 if is_invisible_watermark_available():
@@ -346,11 +346,13 @@ def rescale_noise_cfg(noise_cfg, noise_pred_text, guidance_rescale=0.0):
 
 
 class TextToVideoZeroSDXLPipeline(
+    DeprecatedPipelineMixin,
     DiffusionPipeline,
     StableDiffusionMixin,
     StableDiffusionXLLoraLoaderMixin,
     TextualInversionLoaderMixin,
 ):
+    _last_supported_version = "0.33.1"
     r"""
     Pipeline for zero-shot text-to-video generation using Stable Diffusion XL.
 

src/diffusers/pipelines/unclip/pipeline_unclip.py

@@ -24,7 +24,7 @@ from ...models import PriorTransformer, UNet2DConditionModel, UNet2DModel
 from ...schedulers import UnCLIPScheduler
 from ...utils import is_torch_xla_available, logging
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, ImagePipelineOutput
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, ImagePipelineOutput
 from .text_proj import UnCLIPTextProjModel
 
 
@@ -38,7 +38,7 @@ else:
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 
 
-class UnCLIPPipeline(DiffusionPipeline):
+class UnCLIPPipeline(DeprecatedPipelineMixin, DiffusionPipeline):
     """
     Pipeline for text-to-image generation using unCLIP.
 
@@ -69,6 +69,7 @@ class UnCLIPPipeline(DiffusionPipeline):
 
     """
 
+    _last_supported_version = "0.33.1"
     _exclude_from_cpu_offload = ["prior"]
 
     prior: PriorTransformer

src/diffusers/pipelines/unclip/pipeline_unclip_image_variation.py

@@ -29,7 +29,7 @@ from ...models import UNet2DConditionModel, UNet2DModel
 from ...schedulers import UnCLIPScheduler
 from ...utils import is_torch_xla_available, logging
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, ImagePipelineOutput
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, ImagePipelineOutput
 from .text_proj import UnCLIPTextProjModel
 
 
@@ -43,7 +43,7 @@ else:
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 
 
-class UnCLIPImageVariationPipeline(DiffusionPipeline):
+class UnCLIPImageVariationPipeline(DeprecatedPipelineMixin, DiffusionPipeline):
     """
     Pipeline to generate image variations from an input image using UnCLIP.
 
@@ -73,6 +73,7 @@ class UnCLIPImageVariationPipeline(DiffusionPipeline):
             Scheduler used in the super resolution denoising process (a modified [`DDPMScheduler`]).
     """
 
+    _last_supported_version = "0.33.1"
     decoder: UNet2DConditionModel
     text_proj: UnCLIPTextProjModel
     text_encoder: CLIPTextModelWithProjection

src/diffusers/pipelines/unidiffuser/pipeline_unidiffuser.py

@@ -28,7 +28,7 @@ from ...utils import (
 )
 from ...utils.outputs import BaseOutput
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline
 from .modeling_text_decoder import UniDiffuserTextDecoder
 from .modeling_uvit import UniDiffuserModel
 
@@ -62,7 +62,7 @@ class ImageTextPipelineOutput(BaseOutput):
     text: Optional[Union[List[str], List[List[str]]]]
 
 
-class UniDiffuserPipeline(DiffusionPipeline):
+class UniDiffuserPipeline(DeprecatedPipelineMixin, DiffusionPipeline):
     r"""
     Pipeline for a bimodal image-text model which supports unconditional text and image generation, text-conditioned
     image generation, image-conditioned text generation, and joint image-text generation.
@@ -96,6 +96,7 @@ class UniDiffuserPipeline(DiffusionPipeline):
             original UniDiffuser paper uses the [`DPMSolverMultistepScheduler`] scheduler.
     """
 
+    _last_supported_version = "0.33.1"
     # TODO: support for moving submodules for components with enable_model_cpu_offload
     model_cpu_offload_seq = "text_encoder->image_encoder->unet->vae->text_decoder"
 

src/diffusers/pipelines/wuerstchen/pipeline_wuerstchen.py

@@ -21,7 +21,7 @@ from transformers import CLIPTextModel, CLIPTokenizer
 from ...schedulers import DDPMWuerstchenScheduler
 from ...utils import deprecate, is_torch_xla_available, logging, replace_example_docstring
 from ...utils.torch_utils import randn_tensor
-from ..pipeline_utils import DiffusionPipeline, ImagePipelineOutput
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline, ImagePipelineOutput
 from .modeling_paella_vq_model import PaellaVQModel
 from .modeling_wuerstchen_diffnext import WuerstchenDiffNeXt
 
@@ -56,7 +56,7 @@ EXAMPLE_DOC_STRING = """
 """
 
 
-class WuerstchenDecoderPipeline(DiffusionPipeline):
+class WuerstchenDecoderPipeline(DeprecatedPipelineMixin, DiffusionPipeline):
     """
     Pipeline for generating images from the Wuerstchen model.
 

src/diffusers/pipelines/wuerstchen/pipeline_wuerstchen_combined.py

@@ -18,7 +18,7 @@ from transformers import CLIPTextModel, CLIPTokenizer
 
 from ...schedulers import DDPMWuerstchenScheduler
 from ...utils import deprecate, replace_example_docstring
-from ..pipeline_utils import DiffusionPipeline
+from ..pipeline_utils import DeprecatedPipelineMixin, DiffusionPipeline
 from .modeling_paella_vq_model import PaellaVQModel
 from .modeling_wuerstchen_diffnext import WuerstchenDiffNeXt
 from .modeling_wuerstchen_prior import WuerstchenPrior
@@ -40,7 +40,7 @@ TEXT2IMAGE_EXAMPLE_DOC_STRING = """
 """
 
 
-class WuerstchenCombinedPipeline(DiffusionPipeline):
+class WuerstchenCombinedPipeline(DeprecatedPipelineMixin, DiffusionPipeline):
     """
     Combined Pipeline for text-to-image generation using Wuerstchen
 
@@ -68,6 +68,7 @@ class WuerstchenCombinedPipeline(DiffusionPipeline):
             The scheduler to be used for prior pipeline.
     """
 
+    _last_supported_version = "0.33.1"
     _load_connected_pipes = True
 
     def __init__(

tests/pipelines/test_pipelines_common.py

@@ -1115,6 +1115,14 @@ class PipelineTesterMixin:
         gc.collect()
         backend_empty_cache(torch_device)
 
+        # Skip tests for pipelines that inherit from DeprecatedPipelineMixin
+        from diffusers.pipelines.pipeline_utils import DeprecatedPipelineMixin
+
+        if hasattr(self, "pipeline_class") and issubclass(self.pipeline_class, DeprecatedPipelineMixin):
+            import pytest
+
+            pytest.skip(reason=f"Deprecated Pipeline: {self.pipeline_class.__name__}")
+
     def tearDown(self):
         # clean up the VRAM after each test in case of CUDA runtime errors
         super().tearDown()