diffusers 0.29.2__py3-none-any.whl → 0.30.0__py3-none-any.whl

diffusers/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.29.2"
+__version__ = "0.30.0"
 
 from typing import TYPE_CHECKING
 
@@ -12,6 +12,7 @@ from .utils import (
     is_note_seq_available,
     is_onnx_available,
     is_scipy_available,
+    is_sentencepiece_available,
     is_torch_available,
     is_torchsde_available,
     is_transformers_available,
@@ -76,16 +77,25 @@ else:
     _import_structure["models"].extend(
         [
             "AsymmetricAutoencoderKL",
+            "AuraFlowTransformer2DModel",
             "AutoencoderKL",
+            "AutoencoderKLCogVideoX",
             "AutoencoderKLTemporalDecoder",
+            "AutoencoderOobleck",
             "AutoencoderTiny",
+            "CogVideoXTransformer3DModel",
             "ConsistencyDecoderVAE",
             "ControlNetModel",
             "ControlNetXSAdapter",
             "DiTTransformer2DModel",
+            "FluxTransformer2DModel",
+            "HunyuanDiT2DControlNetModel",
             "HunyuanDiT2DModel",
+            "HunyuanDiT2DMultiControlNetModel",
             "I2VGenXLUNet",
             "Kandinsky3UNet",
+            "LatteTransformer3DModel",
+            "LuminaNextDiT2DModel",
             "ModelMixin",
             "MotionAdapter",
             "MultiAdapter",
@@ -94,6 +104,8 @@ else:
             "SD3ControlNetModel",
             "SD3MultiControlNetModel",
             "SD3Transformer2DModel",
+            "SparseControlNetModel",
+            "StableAudioDiTModel",
             "StableCascadeUNet",
             "T2IAdapter",
             "T5FilmDecoder",
@@ -145,6 +157,8 @@ else:
         [
             "AmusedScheduler",
             "CMStochasticIterativeScheduler",
+            "CogVideoXDDIMScheduler",
+            "CogVideoXDPMScheduler",
             "DDIMInverseScheduler",
             "DDIMParallelScheduler",
             "DDIMScheduler",
@@ -160,6 +174,7 @@ else:
             "EulerAncestralDiscreteScheduler",
             "EulerDiscreteScheduler",
             "FlowMatchEulerDiscreteScheduler",
+            "FlowMatchHeunDiscreteScheduler",
             "HeunDiscreteScheduler",
             "IPNDMScheduler",
             "KarrasVeScheduler",
@@ -203,7 +218,7 @@ except OptionalDependencyNotAvailable:
     ]
 
 else:
-    _import_structure["schedulers"].extend(["DPMSolverSDEScheduler"])
+    _import_structure["schedulers"].extend(["CosineDPMSolverMultistepScheduler", "DPMSolverSDEScheduler"])
 
 try:
     if not (is_torch_available() and is_transformers_available()):
@@ -223,17 +238,25 @@ else:
             "AmusedImg2ImgPipeline",
             "AmusedInpaintPipeline",
             "AmusedPipeline",
+            "AnimateDiffControlNetPipeline",
+            "AnimateDiffPAGPipeline",
             "AnimateDiffPipeline",
             "AnimateDiffSDXLPipeline",
+            "AnimateDiffSparseControlNetPipeline",
             "AnimateDiffVideoToVideoPipeline",
             "AudioLDM2Pipeline",
             "AudioLDM2ProjectionModel",
             "AudioLDM2UNet2DConditionModel",
             "AudioLDMPipeline",
+            "AuraFlowPipeline",
             "BlipDiffusionControlNetPipeline",
             "BlipDiffusionPipeline",
             "CLIPImageProjection",
+            "CogVideoXPipeline",
             "CycleDiffusionPipeline",
+            "FluxPipeline",
+            "HunyuanDiTControlNetPipeline",
+            "HunyuanDiTPAGPipeline",
             "HunyuanDiTPipeline",
             "I2VGenXLPipeline",
             "IFImg2ImgPipeline",
@@ -264,29 +287,37 @@ else:
             "KandinskyV22PriorPipeline",
             "LatentConsistencyModelImg2ImgPipeline",
             "LatentConsistencyModelPipeline",
+            "LattePipeline",
             "LDMTextToImagePipeline",
             "LEditsPPPipelineStableDiffusion",
             "LEditsPPPipelineStableDiffusionXL",
+            "LuminaText2ImgPipeline",
             "MarigoldDepthPipeline",
             "MarigoldNormalsPipeline",
             "MusicLDMPipeline",
             "PaintByExamplePipeline",
             "PIAPipeline",
             "PixArtAlphaPipeline",
+            "PixArtSigmaPAGPipeline",
             "PixArtSigmaPipeline",
             "SemanticStableDiffusionPipeline",
             "ShapEImg2ImgPipeline",
             "ShapEPipeline",
+            "StableAudioPipeline",
+            "StableAudioProjectionModel",
             "StableCascadeCombinedPipeline",
             "StableCascadeDecoderPipeline",
             "StableCascadePriorPipeline",
             "StableDiffusion3ControlNetPipeline",
             "StableDiffusion3Img2ImgPipeline",
+            "StableDiffusion3InpaintPipeline",
+            "StableDiffusion3PAGPipeline",
             "StableDiffusion3Pipeline",
             "StableDiffusionAdapterPipeline",
             "StableDiffusionAttendAndExcitePipeline",
             "StableDiffusionControlNetImg2ImgPipeline",
             "StableDiffusionControlNetInpaintPipeline",
+            "StableDiffusionControlNetPAGPipeline",
             "StableDiffusionControlNetPipeline",
             "StableDiffusionControlNetXSPipeline",
             "StableDiffusionDepth2ImgPipeline",
@@ -301,6 +332,7 @@ else:
             "StableDiffusionLatentUpscalePipeline",
             "StableDiffusionLDM3DPipeline",
             "StableDiffusionModelEditingPipeline",
+            "StableDiffusionPAGPipeline",
             "StableDiffusionPanoramaPipeline",
             "StableDiffusionParadigmsPipeline",
             "StableDiffusionPipeline",
@@ -311,11 +343,15 @@ else:
             "StableDiffusionXLAdapterPipeline",
             "StableDiffusionXLControlNetImg2ImgPipeline",
             "StableDiffusionXLControlNetInpaintPipeline",
+            "StableDiffusionXLControlNetPAGPipeline",
             "StableDiffusionXLControlNetPipeline",
             "StableDiffusionXLControlNetXSPipeline",
             "StableDiffusionXLImg2ImgPipeline",
             "StableDiffusionXLInpaintPipeline",
             "StableDiffusionXLInstructPix2PixPipeline",
+            "StableDiffusionXLPAGImg2ImgPipeline",
+            "StableDiffusionXLPAGInpaintPipeline",
+            "StableDiffusionXLPAGPipeline",
             "StableDiffusionXLPipeline",
             "StableUnCLIPImg2ImgPipeline",
             "StableUnCLIPPipeline",
@@ -353,6 +389,19 @@ except OptionalDependencyNotAvailable:
 else:
     _import_structure["pipelines"].extend(["StableDiffusionKDiffusionPipeline", "StableDiffusionXLKDiffusionPipeline"])
 
+try:
+    if not (is_torch_available() and is_transformers_available() and is_sentencepiece_available()):
+        raise OptionalDependencyNotAvailable()
+except OptionalDependencyNotAvailable:
+    from .utils import dummy_torch_and_transformers_and_sentencepiece_objects  # noqa F403
+
+    _import_structure["utils.dummy_torch_and_transformers_and_sentencepiece_objects"] = [
+        name for name in dir(dummy_torch_and_transformers_and_sentencepiece_objects) if not name.startswith("_")
+    ]
+
+else:
+    _import_structure["pipelines"].extend(["KolorsImg2ImgPipeline", "KolorsPAGPipeline", "KolorsPipeline"])
+
 try:
     if not (is_torch_available() and is_transformers_available() and is_onnx_available()):
         raise OptionalDependencyNotAvailable()
@@ -489,16 +538,25 @@ if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
     else:
         from .models import (
             AsymmetricAutoencoderKL,
+            AuraFlowTransformer2DModel,
             AutoencoderKL,
+            AutoencoderKLCogVideoX,
             AutoencoderKLTemporalDecoder,
+            AutoencoderOobleck,
             AutoencoderTiny,
+            CogVideoXTransformer3DModel,
             ConsistencyDecoderVAE,
             ControlNetModel,
             ControlNetXSAdapter,
             DiTTransformer2DModel,
+            FluxTransformer2DModel,
+            HunyuanDiT2DControlNetModel,
             HunyuanDiT2DModel,
+            HunyuanDiT2DMultiControlNetModel,
             I2VGenXLUNet,
             Kandinsky3UNet,
+            LatteTransformer3DModel,
+            LuminaNextDiT2DModel,
             ModelMixin,
             MotionAdapter,
             MultiAdapter,
@@ -507,6 +565,8 @@ if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
             SD3ControlNetModel,
             SD3MultiControlNetModel,
             SD3Transformer2DModel,
+            SparseControlNetModel,
+            StableAudioDiTModel,
             T2IAdapter,
             T5FilmDecoder,
             Transformer2DModel,
@@ -555,6 +615,8 @@ if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
         from .schedulers import (
             AmusedScheduler,
             CMStochasticIterativeScheduler,
+            CogVideoXDDIMScheduler,
+            CogVideoXDPMScheduler,
             DDIMInverseScheduler,
             DDIMParallelScheduler,
             DDIMScheduler,
@@ -570,6 +632,7 @@ if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
             EulerAncestralDiscreteScheduler,
             EulerDiscreteScheduler,
             FlowMatchEulerDiscreteScheduler,
+            FlowMatchHeunDiscreteScheduler,
             HeunDiscreteScheduler,
             IPNDMScheduler,
             KarrasVeScheduler,
@@ -602,7 +665,7 @@ if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
     except OptionalDependencyNotAvailable:
         from .utils.dummy_torch_and_torchsde_objects import *  # noqa F403
     else:
-        from .schedulers import DPMSolverSDEScheduler
+        from .schedulers import CosineDPMSolverMultistepScheduler, DPMSolverSDEScheduler
 
     try:
         if not (is_torch_available() and is_transformers_available()):
@@ -616,15 +679,23 @@ if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
             AmusedImg2ImgPipeline,
             AmusedInpaintPipeline,
             AmusedPipeline,
+            AnimateDiffControlNetPipeline,
+            AnimateDiffPAGPipeline,
             AnimateDiffPipeline,
             AnimateDiffSDXLPipeline,
+            AnimateDiffSparseControlNetPipeline,
             AnimateDiffVideoToVideoPipeline,
             AudioLDM2Pipeline,
             AudioLDM2ProjectionModel,
             AudioLDM2UNet2DConditionModel,
             AudioLDMPipeline,
+            AuraFlowPipeline,
             CLIPImageProjection,
+            CogVideoXPipeline,
             CycleDiffusionPipeline,
+            FluxPipeline,
+            HunyuanDiTControlNetPipeline,
+            HunyuanDiTPAGPipeline,
             HunyuanDiTPipeline,
             I2VGenXLPipeline,
             IFImg2ImgPipeline,
@@ -655,29 +726,37 @@ if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
             KandinskyV22PriorPipeline,
             LatentConsistencyModelImg2ImgPipeline,
             LatentConsistencyModelPipeline,
+            LattePipeline,
             LDMTextToImagePipeline,
             LEditsPPPipelineStableDiffusion,
             LEditsPPPipelineStableDiffusionXL,
+            LuminaText2ImgPipeline,
             MarigoldDepthPipeline,
             MarigoldNormalsPipeline,
             MusicLDMPipeline,
             PaintByExamplePipeline,
             PIAPipeline,
             PixArtAlphaPipeline,
+            PixArtSigmaPAGPipeline,
             PixArtSigmaPipeline,
             SemanticStableDiffusionPipeline,
             ShapEImg2ImgPipeline,
             ShapEPipeline,
+            StableAudioPipeline,
+            StableAudioProjectionModel,
             StableCascadeCombinedPipeline,
             StableCascadeDecoderPipeline,
             StableCascadePriorPipeline,
             StableDiffusion3ControlNetPipeline,
             StableDiffusion3Img2ImgPipeline,
+            StableDiffusion3InpaintPipeline,
+            StableDiffusion3PAGPipeline,
             StableDiffusion3Pipeline,
             StableDiffusionAdapterPipeline,
             StableDiffusionAttendAndExcitePipeline,
             StableDiffusionControlNetImg2ImgPipeline,
             StableDiffusionControlNetInpaintPipeline,
+            StableDiffusionControlNetPAGPipeline,
             StableDiffusionControlNetPipeline,
             StableDiffusionControlNetXSPipeline,
             StableDiffusionDepth2ImgPipeline,
@@ -692,6 +771,7 @@ if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
             StableDiffusionLatentUpscalePipeline,
             StableDiffusionLDM3DPipeline,
             StableDiffusionModelEditingPipeline,
+            StableDiffusionPAGPipeline,
             StableDiffusionPanoramaPipeline,
             StableDiffusionParadigmsPipeline,
             StableDiffusionPipeline,
@@ -702,11 +782,15 @@ if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
             StableDiffusionXLAdapterPipeline,
             StableDiffusionXLControlNetImg2ImgPipeline,
             StableDiffusionXLControlNetInpaintPipeline,
+            StableDiffusionXLControlNetPAGPipeline,
             StableDiffusionXLControlNetPipeline,
             StableDiffusionXLControlNetXSPipeline,
             StableDiffusionXLImg2ImgPipeline,
             StableDiffusionXLInpaintPipeline,
             StableDiffusionXLInstructPix2PixPipeline,
+            StableDiffusionXLPAGImg2ImgPipeline,
+            StableDiffusionXLPAGInpaintPipeline,
+            StableDiffusionXLPAGPipeline,
             StableDiffusionXLPipeline,
             StableUnCLIPImg2ImgPipeline,
             StableUnCLIPPipeline,
@@ -738,6 +822,13 @@ if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
     else:
         from .pipelines import StableDiffusionKDiffusionPipeline, StableDiffusionXLKDiffusionPipeline
 
+    try:
+        if not (is_torch_available() and is_transformers_available() and is_sentencepiece_available()):
+            raise OptionalDependencyNotAvailable()
+    except OptionalDependencyNotAvailable:
+        from .utils.dummy_torch_and_transformers_and_sentencepiece_objects import *  # noqa F403
+    else:
+        from .pipelines import KolorsImg2ImgPipeline, KolorsPAGPipeline, KolorsPipeline
     try:
         if not (is_torch_available() and is_transformers_available() and is_onnx_available()):
             raise OptionalDependencyNotAvailable()

diffusers/commands/env.py

@@ -24,7 +24,6 @@ from ..utils import (
     is_bitsandbytes_available,
     is_flax_available,
     is_google_colab,
-    is_notebook,
     is_peft_available,
     is_safetensors_available,
     is_torch_available,
@@ -107,8 +106,6 @@ class EnvironmentCommand(BaseDiffusersCLICommand):
 
         platform_info = platform.platform()
 
-        is_notebook_str = "Yes" if is_notebook() else "No"
-
         is_google_colab_str = "Yes" if is_google_colab() else "No"
 
         accelerator = "NA"
@@ -123,7 +120,7 @@ class EnvironmentCommand(BaseDiffusersCLICommand):
                 out_str = out_str.decode("utf-8")
 
                 if len(out_str) > 0:
-                    accelerator = out_str.strip() + " VRAM"
+                    accelerator = out_str.strip()
             except FileNotFoundError:
                 pass
         elif platform.system() == "Darwin":  # Mac OS
@@ -155,7 +152,6 @@ class EnvironmentCommand(BaseDiffusersCLICommand):
         info = {
             "🤗 Diffusers version": version,
             "Platform": platform_info,
-            "Running on a notebook?": is_notebook_str,
             "Running on Google Colab?": is_google_colab_str,
             "Python version": platform.python_version(),
             "PyTorch version (GPU?)": f"{pt_version} ({pt_cuda_available})",

diffusers/configuration_utils.py

@@ -23,7 +23,7 @@ import json
 import os
 import re
 from collections import OrderedDict
-from pathlib import PosixPath
+from pathlib import Path
 from typing import Any, Dict, Tuple, Union
 
 import numpy as np
@@ -310,9 +310,6 @@ class ConfigMixin:
             force_download (`bool`, *optional*, defaults to `False`):
                 Whether or not to force the (re-)download of the model weights and configuration files, overriding the
                 cached versions if they exist.
-            resume_download:
-                Deprecated and ignored. All downloads are now resumed by default when possible. Will be removed in v1
-                of Diffusers.
             proxies (`Dict[str, str]`, *optional*):
                 A dictionary of proxy servers to use by protocol or endpoint, for example, `{'http': 'foo.bar:3128',
                 'http://hostname': 'foo.bar:4012'}`. The proxies are used on each request.
@@ -343,7 +340,6 @@ class ConfigMixin:
         local_dir = kwargs.pop("local_dir", None)
         local_dir_use_symlinks = kwargs.pop("local_dir_use_symlinks", "auto")
         force_download = kwargs.pop("force_download", False)
-        resume_download = kwargs.pop("resume_download", None)
         proxies = kwargs.pop("proxies", None)
         token = kwargs.pop("token", None)
         local_files_only = kwargs.pop("local_files_only", False)
@@ -386,7 +382,6 @@ class ConfigMixin:
                     cache_dir=cache_dir,
                     force_download=force_download,
                     proxies=proxies,
-                    resume_download=resume_download,
                     local_files_only=local_files_only,
                     token=token,
                     user_agent=user_agent,
@@ -587,8 +582,8 @@ class ConfigMixin:
         def to_json_saveable(value):
             if isinstance(value, np.ndarray):
                 value = value.tolist()
-            elif isinstance(value, PosixPath):
-                value = str(value)
+            elif isinstance(value, Path):
+                value = value.as_posix()
             return value
 
         config_dict = {k: to_json_saveable(v) for k, v in config_dict.items()}
@@ -716,7 +711,7 @@ class LegacyConfigMixin(ConfigMixin):
 
     @classmethod
     def from_config(cls, config: Union[FrozenDict, Dict[str, Any]] = None, return_unused_kwargs=False, **kwargs):
-        # To prevent depedency import problem.
+        # To prevent dependency import problem.
         from .models.model_loading_utils import _fetch_remapped_cls_from_config
 
         # resolve remapping

diffusers/dependency_versions_table.py

@@ -3,7 +3,7 @@
 # 2. run `make deps_table_update`
 deps = {
     "Pillow": "Pillow",
-    "accelerate": "accelerate>=0.29.3",
+    "accelerate": "accelerate>=0.31.0",
     "compel": "compel==0.1.8",
     "datasets": "datasets",
     "filelock": "filelock",
@@ -40,7 +40,7 @@ deps = {
     "tensorboard": "tensorboard",
     "torch": "torch>=1.4",
     "torchvision": "torchvision",
-    "transformers": "transformers>=4.25.1",
+    "transformers": "transformers>=4.41.2",
     "urllib3": "urllib3<=2.0.0",
     "black": "black",
 }

diffusers/image_processor.py

@@ -569,7 +569,7 @@ class VaeImageProcessor(ConfigMixin):
 
             channel = image.shape[1]
             # don't need any preprocess if the image is latents
-            if channel == 4:
+            if channel == self.vae_latent_channels:
                 return image
 
             height, width = self.get_default_height_width(image, height, width)
@@ -585,7 +585,6 @@ class VaeImageProcessor(ConfigMixin):
                 FutureWarning,
             )
             do_normalize = False
-
         if do_normalize:
             image = self.normalize(image)
 

diffusers/loaders/__init__.py

@@ -55,11 +55,19 @@ _import_structure = {}
 
 if is_torch_available():
     _import_structure["single_file_model"] = ["FromOriginalModelMixin"]
+
     _import_structure["unet"] = ["UNet2DConditionLoadersMixin"]
     _import_structure["utils"] = ["AttnProcsLayers"]
     if is_transformers_available():
         _import_structure["single_file"] = ["FromSingleFileMixin"]
-        _import_structure["lora"] = ["LoraLoaderMixin", "StableDiffusionXLLoraLoaderMixin", "SD3LoraLoaderMixin"]
+        _import_structure["lora_pipeline"] = [
+            "AmusedLoraLoaderMixin",
+            "StableDiffusionLoraLoaderMixin",
+            "SD3LoraLoaderMixin",
+            "StableDiffusionXLLoraLoaderMixin",
+            "LoraLoaderMixin",
+            "FluxLoraLoaderMixin",
+        ]
         _import_structure["textual_inversion"] = ["TextualInversionLoaderMixin"]
         _import_structure["ip_adapter"] = ["IPAdapterMixin"]
 
@@ -74,7 +82,14 @@ if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
 
         if is_transformers_available():
             from .ip_adapter import IPAdapterMixin
-            from .lora import LoraLoaderMixin, SD3LoraLoaderMixin, StableDiffusionXLLoraLoaderMixin
+            from .lora_pipeline import (
+                AmusedLoraLoaderMixin,
+                FluxLoraLoaderMixin,
+                LoraLoaderMixin,
+                SD3LoraLoaderMixin,
+                StableDiffusionLoraLoaderMixin,
+                StableDiffusionXLLoraLoaderMixin,
+            )
             from .single_file import FromSingleFileMixin
             from .textual_inversion import TextualInversionLoaderMixin
 

diffusers/loaders/ip_adapter.py

@@ -90,9 +90,7 @@ class IPAdapterMixin:
             force_download (`bool`, *optional*, defaults to `False`):
                 Whether or not to force the (re-)download of the model weights and configuration files, overriding the
                 cached versions if they exist.
-            resume_download:
-                Deprecated and ignored. All downloads are now resumed by default when possible. Will be removed in v1
-                of Diffusers.
+
             proxies (`Dict[str, str]`, *optional*):
                 A dictionary of proxy servers to use by protocol or endpoint, for example, `{'http': 'foo.bar:3128',
                 'http://hostname': 'foo.bar:4012'}`. The proxies are used on each request.
@@ -135,7 +133,6 @@ class IPAdapterMixin:
         # Load the main state dict first.
         cache_dir = kwargs.pop("cache_dir", None)
         force_download = kwargs.pop("force_download", False)
-        resume_download = kwargs.pop("resume_download", None)
         proxies = kwargs.pop("proxies", None)
         local_files_only = kwargs.pop("local_files_only", None)
         token = kwargs.pop("token", None)
@@ -171,7 +168,6 @@ class IPAdapterMixin:
                     weights_name=weight_name,
                     cache_dir=cache_dir,
                     force_download=force_download,
-                    resume_download=resume_download,
                     proxies=proxies,
                     local_files_only=local_files_only,
                     token=token,
@@ -226,7 +222,8 @@ class IPAdapterMixin:
 
             # create feature extractor if it has not been registered to the pipeline yet
             if hasattr(self, "feature_extractor") and getattr(self, "feature_extractor", None) is None:
-                feature_extractor = CLIPImageProcessor()
+                clip_image_size = self.image_encoder.config.image_size
+                feature_extractor = CLIPImageProcessor(size=clip_image_size, crop_size=clip_image_size)
                 self.register_modules(feature_extractor=feature_extractor)
 
         # load ip-adapter into unet
@@ -323,7 +320,13 @@ class IPAdapterMixin:
 
         # remove hidden encoder
         self.unet.encoder_hid_proj = None
-        self.config.encoder_hid_dim_type = None
+        self.unet.config.encoder_hid_dim_type = None
+
+        # Kolors: restore `encoder_hid_proj` with `text_encoder_hid_proj`
+        if hasattr(self.unet, "text_encoder_hid_proj") and self.unet.text_encoder_hid_proj is not None:
+            self.unet.encoder_hid_proj = self.unet.text_encoder_hid_proj
+            self.unet.text_encoder_hid_proj = None
+            self.unet.config.encoder_hid_dim_type = "text_proj"
 
         # restore original Unet attention processors layers
         attn_procs = {}