Merge branch 'main' into group-offloading-with-disk

Author: sayakpaul

src/diffusers/pipelines/amused/pipeline_amused.py

@@ -21,7 +21,7 @@
 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 @@
 """
 
 
-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 ...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 @@
 """
 
 
-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 ...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 @@
 """
 
 
-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 ...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 @@
 """
 
 
-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 @@
     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 @@
 """
 
 
-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 @@
     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 @@
 
 
 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 @@
     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 @@
 
 
 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 ...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 @@
 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.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 @@
     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 @@
 """
 
 
-class MusicLDMPipeline(DiffusionPipeline, StableDiffusionMixin):
+class MusicLDMPipeline(DeprecatedPipelineMixin, DiffusionPipeline, StableDiffusionMixin):
+    _last_supported_version = "0.33.1"
     r"""
     Pipeline for text-to-audio generation using MusicLDM.