diffusers 0.32.2__py3-none-any.whl → 0.33.0__py3-none-any.whl

diffusers/pipelines/semantic_stable_diffusion/pipeline_semantic_stable_diffusion.py

@@ -9,12 +9,19 @@ from ...image_processor import VaeImageProcessor
 from ...models import AutoencoderKL, UNet2DConditionModel
 from ...pipelines.stable_diffusion.safety_checker import StableDiffusionSafetyChecker
 from ...schedulers import KarrasDiffusionSchedulers
-from ...utils import deprecate, logging
+from ...utils import deprecate, is_torch_xla_available, logging
 from ...utils.torch_utils import randn_tensor
 from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
 from .pipeline_output import SemanticStableDiffusionPipelineOutput
 
 
+if is_torch_xla_available():
+    import torch_xla.core.xla_model as xm
+
+    XLA_AVAILABLE = True
+else:
+    XLA_AVAILABLE = False
+
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 
 
@@ -87,7 +94,7 @@ class SemanticStableDiffusionPipeline(DiffusionPipeline, StableDiffusionMixin):
             safety_checker=safety_checker,
             feature_extractor=feature_extractor,
         )
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
         self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
         self.register_to_config(requires_safety_checker=requires_safety_checker)
 
@@ -701,6 +708,9 @@ class SemanticStableDiffusionPipeline(DiffusionPipeline, StableDiffusionMixin):
                 step_idx = i // getattr(self.scheduler, "order", 1)
                 callback(step_idx, t, latents)
 
+            if XLA_AVAILABLE:
+                xm.mark_step()
+
         # 8. Post-processing
         if not output_type == "latent":
             image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False)[0]

diffusers/pipelines/shap_e/pipeline_shap_e.py

@@ -25,6 +25,7 @@ from ...models import PriorTransformer
 from ...schedulers import HeunDiscreteScheduler
 from ...utils import (
     BaseOutput,
+    is_torch_xla_available,
     logging,
     replace_example_docstring,
 )
@@ -33,8 +34,16 @@ from ..pipeline_utils import DiffusionPipeline
 from .renderer import ShapERenderer
 
 
+if is_torch_xla_available():
+    import torch_xla.core.xla_model as xm
+
+    XLA_AVAILABLE = True
+else:
+    XLA_AVAILABLE = False
+
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 
+
 EXAMPLE_DOC_STRING = """
     Examples:
         ```py
@@ -291,6 +300,9 @@ class ShapEPipeline(DiffusionPipeline):
                 sample=latents,
             ).prev_sample
 
+            if XLA_AVAILABLE:
+                xm.mark_step()
+
         # Offload all models
         self.maybe_free_model_hooks()
 

diffusers/pipelines/shap_e/pipeline_shap_e_img2img.py

@@ -24,6 +24,7 @@ from ...models import PriorTransformer
 from ...schedulers import HeunDiscreteScheduler
 from ...utils import (
     BaseOutput,
+    is_torch_xla_available,
     logging,
     replace_example_docstring,
 )
@@ -32,8 +33,16 @@ from ..pipeline_utils import DiffusionPipeline
 from .renderer import ShapERenderer
 
 
+if is_torch_xla_available():
+    import torch_xla.core.xla_model as xm
+
+    XLA_AVAILABLE = True
+else:
+    XLA_AVAILABLE = False
+
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 
+
 EXAMPLE_DOC_STRING = """
     Examples:
         ```py
@@ -278,6 +287,9 @@ class ShapEImg2ImgPipeline(DiffusionPipeline):
                 sample=latents,
             ).prev_sample
 
+            if XLA_AVAILABLE:
+                xm.mark_step()
+
         if output_type not in ["np", "pil", "latent", "mesh"]:
             raise ValueError(
                 f"Only the output types `pil`, `np`, `latent` and `mesh` are supported not output_type={output_type}"

diffusers/pipelines/shap_e/renderer.py

@@ -983,9 +983,9 @@ class ShapERenderer(ModelMixin, ConfigMixin):
         fields = torch.cat(fields, dim=1)
         fields = fields.float()
 
-        assert (
-            len(fields.shape) == 3 and fields.shape[-1] == 1
-        ), f"expected [meta_batch x inner_batch] SDF results, but got {fields.shape}"
+        assert len(fields.shape) == 3 and fields.shape[-1] == 1, (
+            f"expected [meta_batch x inner_batch] SDF results, but got {fields.shape}"
+        )
 
         fields = fields.reshape(1, *([grid_size] * 3))
 
@@ -1039,9 +1039,9 @@ class ShapERenderer(ModelMixin, ConfigMixin):
         textures = textures.float()
 
         # 3.3 augument the mesh with texture data
-        assert len(textures.shape) == 3 and textures.shape[-1] == len(
-            texture_channels
-        ), f"expected [meta_batch x inner_batch x texture_channels] field results, but got {textures.shape}"
+        assert len(textures.shape) == 3 and textures.shape[-1] == len(texture_channels), (
+            f"expected [meta_batch x inner_batch x texture_channels] field results, but got {textures.shape}"
+        )
 
         for m, texture in zip(raw_meshes, textures):
             texture = texture[: len(m.verts)]

diffusers/pipelines/stable_audio/pipeline_stable_audio.py

@@ -584,7 +584,7 @@ class StableAudioPipeline(DiffusionPipeline):
 
         if audio_end_in_s - audio_start_in_s > max_audio_length_in_s:
             raise ValueError(
-                f"The total audio length requested ({audio_end_in_s-audio_start_in_s}s) is longer than the model maximum possible length ({max_audio_length_in_s}). Make sure that 'audio_end_in_s-audio_start_in_s<={max_audio_length_in_s}'."
+                f"The total audio length requested ({audio_end_in_s - audio_start_in_s}s) is longer than the model maximum possible length ({max_audio_length_in_s}). Make sure that 'audio_end_in_s-audio_start_in_s<={max_audio_length_in_s}'."
             )
 
         waveform_start = int(audio_start_in_s * self.vae.config.sampling_rate)

diffusers/pipelines/stable_cascade/pipeline_stable_cascade.py

@@ -15,18 +15,26 @@
 from typing import Callable, Dict, List, Optional, Union
 
 import torch
-from transformers import CLIPTextModel, CLIPTokenizer
+from transformers import CLIPTextModelWithProjection, CLIPTokenizer
 
 from ...models import StableCascadeUNet
 from ...schedulers import DDPMWuerstchenScheduler
-from ...utils import is_torch_version, logging, replace_example_docstring
+from ...utils import is_torch_version, is_torch_xla_available, logging, replace_example_docstring
 from ...utils.torch_utils import randn_tensor
 from ..pipeline_utils import DiffusionPipeline, ImagePipelineOutput
 from ..wuerstchen.modeling_paella_vq_model import PaellaVQModel
 
 
+if is_torch_xla_available():
+    import torch_xla.core.xla_model as xm
+
+    XLA_AVAILABLE = True
+else:
+    XLA_AVAILABLE = False
+
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 
+
 EXAMPLE_DOC_STRING = """
     Examples:
         ```py
@@ -57,7 +65,7 @@ class StableCascadeDecoderPipeline(DiffusionPipeline):
     Args:
         tokenizer (`CLIPTokenizer`):
             The CLIP tokenizer.
-        text_encoder (`CLIPTextModel`):
+        text_encoder (`CLIPTextModelWithProjection`):
             The CLIP text encoder.
         decoder ([`StableCascadeUNet`]):
             The Stable Cascade decoder unet.
@@ -85,7 +93,7 @@ class StableCascadeDecoderPipeline(DiffusionPipeline):
         self,
         decoder: StableCascadeUNet,
         tokenizer: CLIPTokenizer,
-        text_encoder: CLIPTextModel,
+        text_encoder: CLIPTextModelWithProjection,
         scheduler: DDPMWuerstchenScheduler,
         vqgan: PaellaVQModel,
         latent_dim_scale: float = 10.67,
@@ -503,6 +511,9 @@ class StableCascadeDecoderPipeline(DiffusionPipeline):
                 prompt_embeds = callback_outputs.pop("prompt_embeds", prompt_embeds)
                 negative_prompt_embeds = callback_outputs.pop("negative_prompt_embeds", negative_prompt_embeds)
 
+            if XLA_AVAILABLE:
+                xm.mark_step()
+
         if output_type not in ["pt", "np", "pil", "latent"]:
             raise ValueError(
                 f"Only the output types `pt`, `np`, `pil` and `latent` are supported not output_type={output_type}"

diffusers/pipelines/stable_cascade/pipeline_stable_cascade_combined.py

@@ -15,7 +15,7 @@ from typing import Callable, Dict, List, Optional, Union
 
 import PIL
 import torch
-from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer, CLIPVisionModelWithProjection
+from transformers import CLIPImageProcessor, CLIPTextModelWithProjection, CLIPTokenizer, CLIPVisionModelWithProjection
 
 from ...models import StableCascadeUNet
 from ...schedulers import DDPMWuerstchenScheduler
@@ -52,7 +52,7 @@ class StableCascadeCombinedPipeline(DiffusionPipeline):
     Args:
         tokenizer (`CLIPTokenizer`):
             The decoder tokenizer to be used for text inputs.
-        text_encoder (`CLIPTextModel`):
+        text_encoder (`CLIPTextModelWithProjection`):
             The decoder text encoder to be used for text inputs.
         decoder (`StableCascadeUNet`):
             The decoder model to be used for decoder image generation pipeline.
@@ -60,14 +60,18 @@ class StableCascadeCombinedPipeline(DiffusionPipeline):
             The scheduler to be used for decoder image generation pipeline.
         vqgan (`PaellaVQModel`):
             The VQGAN model to be used for decoder image generation pipeline.
-        feature_extractor ([`~transformers.CLIPImageProcessor`]):
-            Model that extracts features from generated images to be used as inputs for the `image_encoder`.
-        image_encoder ([`CLIPVisionModelWithProjection`]):
-            Frozen CLIP image-encoder ([clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14)).
         prior_prior (`StableCascadeUNet`):
             The prior model to be used for prior pipeline.
+        prior_text_encoder (`CLIPTextModelWithProjection`):
+            The prior text encoder to be used for text inputs.
+        prior_tokenizer (`CLIPTokenizer`):
+            The prior tokenizer to be used for text inputs.
         prior_scheduler (`DDPMWuerstchenScheduler`):
             The scheduler to be used for prior pipeline.
+        prior_feature_extractor ([`~transformers.CLIPImageProcessor`]):
+            Model that extracts features from generated images to be used as inputs for the `image_encoder`.
+        prior_image_encoder ([`CLIPVisionModelWithProjection`]):
+            Frozen CLIP image-encoder ([clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14)).
     """
 
     _load_connected_pipes = True
@@ -76,12 +80,12 @@ class StableCascadeCombinedPipeline(DiffusionPipeline):
     def __init__(
         self,
         tokenizer: CLIPTokenizer,
-        text_encoder: CLIPTextModel,
+        text_encoder: CLIPTextModelWithProjection,
         decoder: StableCascadeUNet,
         scheduler: DDPMWuerstchenScheduler,
         vqgan: PaellaVQModel,
         prior_prior: StableCascadeUNet,
-        prior_text_encoder: CLIPTextModel,
+        prior_text_encoder: CLIPTextModelWithProjection,
         prior_tokenizer: CLIPTokenizer,
         prior_scheduler: DDPMWuerstchenScheduler,
         prior_feature_extractor: Optional[CLIPImageProcessor] = None,

diffusers/pipelines/stable_cascade/pipeline_stable_cascade_prior.py

@@ -23,13 +23,21 @@ from transformers import CLIPImageProcessor, CLIPTextModelWithProjection, CLIPTo
 
 from ...models import StableCascadeUNet
 from ...schedulers import DDPMWuerstchenScheduler
-from ...utils import BaseOutput, logging, replace_example_docstring
+from ...utils import BaseOutput, is_torch_xla_available, logging, replace_example_docstring
 from ...utils.torch_utils import randn_tensor
 from ..pipeline_utils import DiffusionPipeline
 
 
+if is_torch_xla_available():
+    import torch_xla.core.xla_model as xm
+
+    XLA_AVAILABLE = True
+else:
+    XLA_AVAILABLE = False
+
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 
+
 DEFAULT_STAGE_C_TIMESTEPS = list(np.linspace(1.0, 2 / 3, 20)) + list(np.linspace(2 / 3, 0.0, 11))[1:]
 
 EXAMPLE_DOC_STRING = """
@@ -611,6 +619,9 @@ class StableCascadePriorPipeline(DiffusionPipeline):
                 prompt_embeds = callback_outputs.pop("prompt_embeds", prompt_embeds)
                 negative_prompt_embeds = callback_outputs.pop("negative_prompt_embeds", negative_prompt_embeds)
 
+            if XLA_AVAILABLE:
+                xm.mark_step()
+
         # Offload all models
         self.maybe_free_model_hooks()
 

diffusers/pipelines/stable_diffusion/convert_from_ckpt.py

@@ -1,5 +1,5 @@
 # coding=utf-8
-# Copyright 2024 The HuggingFace Inc. team.
+# Copyright 2025 The HuggingFace Inc. team.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -52,6 +52,7 @@ from ...schedulers import (
     UnCLIPScheduler,
 )
 from ...utils import is_accelerate_available, logging
+from ...utils.constants import DIFFUSERS_REQUEST_TIMEOUT
 from ..latent_diffusion.pipeline_latent_diffusion import LDMBertConfig, LDMBertModel
 from ..paint_by_example import PaintByExampleImageEncoder
 from ..pipeline_utils import DiffusionPipeline
@@ -1324,7 +1325,7 @@ def download_from_original_stable_diffusion_ckpt(
             config_url = "https://raw.githubusercontent.com/Stability-AI/stablediffusion/main/configs/stable-diffusion/x4-upscaling.yaml"
 
         if config_url is not None:
-            original_config_file = BytesIO(requests.get(config_url).content)
+            original_config_file = BytesIO(requests.get(config_url, timeout=DIFFUSERS_REQUEST_TIMEOUT).content)
         else:
             with open(original_config_file, "r") as f:
                 original_config_file = f.read()

diffusers/pipelines/stable_diffusion/pipeline_flax_stable_diffusion.py

@@ -55,7 +55,7 @@ EXAMPLE_DOC_STRING = """
         >>> from diffusers import FlaxStableDiffusionPipeline
 
         >>> pipeline, params = FlaxStableDiffusionPipeline.from_pretrained(
-        ...     "runwayml/stable-diffusion-v1-5", variant="bf16", dtype=jax.numpy.bfloat16
+        ...     "stable-diffusion-v1-5/stable-diffusion-v1-5", variant="bf16", dtype=jax.numpy.bfloat16
         ... )
 
         >>> prompt = "a photo of an astronaut riding a horse on mars"
@@ -100,8 +100,8 @@ class FlaxStableDiffusionPipeline(FlaxDiffusionPipeline):
             [`FlaxDPMSolverMultistepScheduler`].
         safety_checker ([`FlaxStableDiffusionSafetyChecker`]):
             Classification module that estimates whether generated images could be considered offensive or harmful.
-            Please refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for more details
-            about a model's potential harms.
+            Please refer to the [model card](https://huggingface.co/stable-diffusion-v1-5/stable-diffusion-v1-5) for
+            more details about a model's potential harms.
         feature_extractor ([`~transformers.CLIPImageProcessor`]):
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
@@ -132,17 +132,21 @@ class FlaxStableDiffusionPipeline(FlaxDiffusionPipeline):
                 " information, please have a look at https://github.com/huggingface/diffusers/pull/254 ."
             )
 
-        is_unet_version_less_0_9_0 = hasattr(unet.config, "_diffusers_version") and version.parse(
-            version.parse(unet.config._diffusers_version).base_version
-        ) < version.parse("0.9.0.dev0")
-        is_unet_sample_size_less_64 = hasattr(unet.config, "sample_size") and unet.config.sample_size < 64
+        is_unet_version_less_0_9_0 = (
+            unet is not None
+            and hasattr(unet.config, "_diffusers_version")
+            and version.parse(version.parse(unet.config._diffusers_version).base_version) < version.parse("0.9.0.dev0")
+        )
+        is_unet_sample_size_less_64 = (
+            unet is not None and hasattr(unet.config, "sample_size") and unet.config.sample_size < 64
+        )
         if is_unet_version_less_0_9_0 and is_unet_sample_size_less_64:
             deprecation_message = (
                 "The configuration file of the unet has set the default `sample_size` to smaller than"
                 " 64 which seems highly unlikely .If you're checkpoint is a fine-tuned version of any of the"
                 " following: \n- CompVis/stable-diffusion-v1-4 \n- CompVis/stable-diffusion-v1-3 \n-"
-                " CompVis/stable-diffusion-v1-2 \n- CompVis/stable-diffusion-v1-1 \n- runwayml/stable-diffusion-v1-5"
-                " \n- runwayml/stable-diffusion-inpainting \n you should change 'sample_size' to 64 in the"
+                " CompVis/stable-diffusion-v1-2 \n- CompVis/stable-diffusion-v1-1 \n- stable-diffusion-v1-5/stable-diffusion-v1-5"
+                " \n- stable-diffusion-v1-5/stable-diffusion-inpainting \n you should change 'sample_size' to 64 in the"
                 " configuration file. Please make sure to update the config accordingly as leaving `sample_size=32`"
                 " in the config might lead to incorrect results in future versions. If you have downloaded this"
                 " checkpoint from the Hugging Face Hub, it would be very nice if you could open a Pull request for"
@@ -162,7 +166,7 @@ class FlaxStableDiffusionPipeline(FlaxDiffusionPipeline):
             safety_checker=safety_checker,
             feature_extractor=feature_extractor,
         )
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
 
     def prepare_inputs(self, prompt: Union[str, List[str]]):
         if not isinstance(prompt, (str, list)):

diffusers/pipelines/stable_diffusion/pipeline_flax_stable_diffusion_img2img.py

@@ -124,8 +124,8 @@ class FlaxStableDiffusionImg2ImgPipeline(FlaxDiffusionPipeline):
             [`FlaxDPMSolverMultistepScheduler`].
         safety_checker ([`FlaxStableDiffusionSafetyChecker`]):
             Classification module that estimates whether generated images could be considered offensive or harmful.
-            Please refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for more details
-            about a model's potential harms.
+            Please refer to the [model card](https://huggingface.co/stable-diffusion-v1-5/stable-diffusion-v1-5) for
+            more details about a model's potential harms.
         feature_extractor ([`~transformers.CLIPImageProcessor`]):
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
@@ -165,7 +165,7 @@ class FlaxStableDiffusionImg2ImgPipeline(FlaxDiffusionPipeline):
             safety_checker=safety_checker,
             feature_extractor=feature_extractor,
         )
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
 
     def prepare_inputs(self, prompt: Union[str, List[str]], image: Union[Image.Image, List[Image.Image]]):
         if not isinstance(prompt, (str, list)):

diffusers/pipelines/stable_diffusion/pipeline_flax_stable_diffusion_inpaint.py

@@ -127,8 +127,8 @@ class FlaxStableDiffusionInpaintPipeline(FlaxDiffusionPipeline):
             [`FlaxDPMSolverMultistepScheduler`].
         safety_checker ([`FlaxStableDiffusionSafetyChecker`]):
             Classification module that estimates whether generated images could be considered offensive or harmful.
-            Please refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for more details
-            about a model's potential harms.
+            Please refer to the [model card](https://huggingface.co/stable-diffusion-v1-5/stable-diffusion-v1-5) for
+            more details about a model's potential harms.
         feature_extractor ([`~transformers.CLIPImageProcessor`]):
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
@@ -159,17 +159,21 @@ class FlaxStableDiffusionInpaintPipeline(FlaxDiffusionPipeline):
                 " information, please have a look at https://github.com/huggingface/diffusers/pull/254 ."
             )
 
-        is_unet_version_less_0_9_0 = hasattr(unet.config, "_diffusers_version") and version.parse(
-            version.parse(unet.config._diffusers_version).base_version
-        ) < version.parse("0.9.0.dev0")
-        is_unet_sample_size_less_64 = hasattr(unet.config, "sample_size") and unet.config.sample_size < 64
+        is_unet_version_less_0_9_0 = (
+            unet is not None
+            and hasattr(unet.config, "_diffusers_version")
+            and version.parse(version.parse(unet.config._diffusers_version).base_version) < version.parse("0.9.0.dev0")
+        )
+        is_unet_sample_size_less_64 = (
+            unet is not None and hasattr(unet.config, "sample_size") and unet.config.sample_size < 64
+        )
         if is_unet_version_less_0_9_0 and is_unet_sample_size_less_64:
             deprecation_message = (
                 "The configuration file of the unet has set the default `sample_size` to smaller than"
                 " 64 which seems highly unlikely .If you're checkpoint is a fine-tuned version of any of the"
                 " following: \n- CompVis/stable-diffusion-v1-4 \n- CompVis/stable-diffusion-v1-3 \n-"
-                " CompVis/stable-diffusion-v1-2 \n- CompVis/stable-diffusion-v1-1 \n- runwayml/stable-diffusion-v1-5"
-                " \n- runwayml/stable-diffusion-inpainting \n you should change 'sample_size' to 64 in the"
+                " CompVis/stable-diffusion-v1-2 \n- CompVis/stable-diffusion-v1-1 \n- stable-diffusion-v1-5/stable-diffusion-v1-5"
+                " \n- stable-diffusion-v1-5/stable-diffusion-inpainting \n you should change 'sample_size' to 64 in the"
                 " configuration file. Please make sure to update the config accordingly as leaving `sample_size=32`"
                 " in the config might lead to incorrect results in future versions. If you have downloaded this"
                 " checkpoint from the Hugging Face Hub, it would be very nice if you could open a Pull request for"
@@ -189,7 +193,7 @@ class FlaxStableDiffusionInpaintPipeline(FlaxDiffusionPipeline):
             safety_checker=safety_checker,
             feature_extractor=feature_extractor,
         )
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
 
     def prepare_inputs(
         self,
@@ -331,7 +335,7 @@ class FlaxStableDiffusionInpaintPipeline(FlaxDiffusionPipeline):
                 f"Incorrect configuration settings! The config of `pipeline.unet`: {self.unet.config} expects"
                 f" {self.unet.config.in_channels} but received `num_channels_latents`: {num_channels_latents} +"
                 f" `num_channels_mask`: {num_channels_mask} + `num_channels_masked_image`: {num_channels_masked_image}"
-                f" = {num_channels_latents+num_channels_masked_image+num_channels_mask}. Please verify the config of"
+                f" = {num_channels_latents + num_channels_masked_image + num_channels_mask}. Please verify the config of"
                 " `pipeline.unet` or your `mask_image` or `image` input."
             )
 

diffusers/pipelines/stable_diffusion/pipeline_onnx_stable_diffusion.py

@@ -57,7 +57,7 @@ class OnnxStableDiffusionPipeline(DiffusionPipeline):
     ):
         super().__init__()
 
-        if hasattr(scheduler.config, "steps_offset") and scheduler.config.steps_offset != 1:
+        if scheduler is not None and getattr(scheduler.config, "steps_offset", 1) != 1:
             deprecation_message = (
                 f"The configuration file of this scheduler: {scheduler} is outdated. `steps_offset`"
                 f" should be set to 1 instead of {scheduler.config.steps_offset}. Please make sure "
@@ -71,7 +71,7 @@ class OnnxStableDiffusionPipeline(DiffusionPipeline):
             new_config["steps_offset"] = 1
             scheduler._internal_dict = FrozenDict(new_config)
 
-        if hasattr(scheduler.config, "clip_sample") and scheduler.config.clip_sample is True:
+        if scheduler is not None and getattr(scheduler.config, "clip_sample", False) is True:
             deprecation_message = (
                 f"The configuration file of this scheduler: {scheduler} has not set the configuration `clip_sample`."
                 " `clip_sample` should be set to False in the configuration file. Please make sure to update the"

diffusers/pipelines/stable_diffusion/pipeline_onnx_stable_diffusion_img2img.py

@@ -78,7 +78,8 @@ class OnnxStableDiffusionImg2ImgPipeline(DiffusionPipeline):
             [`DDIMScheduler`], [`LMSDiscreteScheduler`], or [`PNDMScheduler`].
         safety_checker ([`StableDiffusionSafetyChecker`]):
             Classification module that estimates whether generated images could be considered offensive or harmful.
-            Please, refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for details.
+            Please, refer to the [model card](https://huggingface.co/stable-diffusion-v1-5/stable-diffusion-v1-5) for
+            details.
         feature_extractor ([`CLIPImageProcessor`]):
             Model that extracts features from generated images to be used as inputs for the `safety_checker`.
     """
@@ -109,7 +110,7 @@ class OnnxStableDiffusionImg2ImgPipeline(DiffusionPipeline):
     ):
         super().__init__()
 
-        if hasattr(scheduler.config, "steps_offset") and scheduler.config.steps_offset != 1:
+        if scheduler is not None and getattr(scheduler.config, "steps_offset", 1) != 1:
             deprecation_message = (
                 f"The configuration file of this scheduler: {scheduler} is outdated. `steps_offset`"
                 f" should be set to 1 instead of {scheduler.config.steps_offset}. Please make sure "
@@ -123,7 +124,7 @@ class OnnxStableDiffusionImg2ImgPipeline(DiffusionPipeline):
             new_config["steps_offset"] = 1
             scheduler._internal_dict = FrozenDict(new_config)
 
-        if hasattr(scheduler.config, "clip_sample") and scheduler.config.clip_sample is True:
+        if scheduler is not None and getattr(scheduler.config, "clip_sample", False) is True:
             deprecation_message = (
                 f"The configuration file of this scheduler: {scheduler} has not set the configuration `clip_sample`."
                 " `clip_sample` should be set to False in the configuration file. Please make sure to update the"

diffusers/pipelines/stable_diffusion/pipeline_onnx_stable_diffusion_inpaint.py

@@ -76,7 +76,8 @@ class OnnxStableDiffusionInpaintPipeline(DiffusionPipeline):
             [`DDIMScheduler`], [`LMSDiscreteScheduler`], or [`PNDMScheduler`].
         safety_checker ([`StableDiffusionSafetyChecker`]):
             Classification module that estimates whether generated images could be considered offensive or harmful.
-            Please, refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for details.
+            Please, refer to the [model card](https://huggingface.co/stable-diffusion-v1-5/stable-diffusion-v1-5) for
+            details.
         feature_extractor ([`CLIPImageProcessor`]):
             Model that extracts features from generated images to be used as inputs for the `safety_checker`.
     """
@@ -108,7 +109,7 @@ class OnnxStableDiffusionInpaintPipeline(DiffusionPipeline):
         super().__init__()
         logger.info("`OnnxStableDiffusionInpaintPipeline` is experimental and will very likely change in the future.")
 
-        if hasattr(scheduler.config, "steps_offset") and scheduler.config.steps_offset != 1:
+        if scheduler is not None and getattr(scheduler.config, "steps_offset", 1) != 1:
             deprecation_message = (
                 f"The configuration file of this scheduler: {scheduler} is outdated. `steps_offset`"
                 f" should be set to 1 instead of {scheduler.config.steps_offset}. Please make sure "
@@ -122,7 +123,7 @@ class OnnxStableDiffusionInpaintPipeline(DiffusionPipeline):
             new_config["steps_offset"] = 1
             scheduler._internal_dict = FrozenDict(new_config)
 
-        if hasattr(scheduler.config, "clip_sample") and scheduler.config.clip_sample is True:
+        if scheduler is not None and getattr(scheduler.config, "clip_sample", False) is True:
             deprecation_message = (
                 f"The configuration file of this scheduler: {scheduler} has not set the configuration `clip_sample`."
                 " `clip_sample` should be set to False in the configuration file. Please make sure to update the"
@@ -474,7 +475,7 @@ class OnnxStableDiffusionInpaintPipeline(DiffusionPipeline):
                 "Incorrect configuration settings! The config of `pipeline.unet` expects"
                 f" {unet_input_channels} but received `num_channels_latents`: {num_channels_latents} +"
                 f" `num_channels_mask`: {num_channels_mask} + `num_channels_masked_image`: {num_channels_masked_image}"
-                f" = {num_channels_latents+num_channels_masked_image+num_channels_mask}. Please verify the config of"
+                f" = {num_channels_latents + num_channels_masked_image + num_channels_mask}. Please verify the config of"
                 " `pipeline.unet` or your `mask_image` or `image` input."
             )
 

diffusers/pipelines/stable_diffusion/pipeline_onnx_stable_diffusion_upscale.py

@@ -83,7 +83,7 @@ class OnnxStableDiffusionUpscalePipeline(DiffusionPipeline):
     ):
         super().__init__()
 
-        if hasattr(scheduler.config, "steps_offset") and scheduler.config.steps_offset != 1:
+        if scheduler is not None and getattr(scheduler.config, "steps_offset", 1) != 1:
             deprecation_message = (
                 f"The configuration file of this scheduler: {scheduler} is outdated. `steps_offset`"
                 f" should be set to 1 instead of {scheduler.config.steps_offset}. Please make sure "
@@ -97,7 +97,7 @@ class OnnxStableDiffusionUpscalePipeline(DiffusionPipeline):
             new_config["steps_offset"] = 1
             scheduler._internal_dict = FrozenDict(new_config)
 
-        if hasattr(scheduler.config, "clip_sample") and scheduler.config.clip_sample is True:
+        if scheduler is not None and getattr(scheduler.config, "clip_sample", False) is True:
             deprecation_message = (
                 f"The configuration file of this scheduler: {scheduler} has not set the configuration `clip_sample`."
                 " `clip_sample` should be set to False in the configuration file. Please make sure to update the"

diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion.py

@@ -55,7 +55,9 @@ EXAMPLE_DOC_STRING = """
         >>> import torch
         >>> from diffusers import StableDiffusionPipeline
 
-        >>> pipe = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16)
+        >>> pipe = StableDiffusionPipeline.from_pretrained(
+        ...     "stable-diffusion-v1-5/stable-diffusion-v1-5", torch_dtype=torch.float16
+        ... )
         >>> pipe = pipe.to("cuda")
 
         >>> prompt = "a photo of an astronaut riding a horse on mars"
@@ -184,8 +186,8 @@ class StableDiffusionPipeline(
             [`DDIMScheduler`], [`LMSDiscreteScheduler`], or [`PNDMScheduler`].
         safety_checker ([`StableDiffusionSafetyChecker`]):
             Classification module that estimates whether generated images could be considered offensive or harmful.
-            Please refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for more details
-            about a model's potential harms.
+            Please refer to the [model card](https://huggingface.co/stable-diffusion-v1-5/stable-diffusion-v1-5) for
+            more details about a model's potential harms.
         feature_extractor ([`~transformers.CLIPImageProcessor`]):
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
@@ -209,7 +211,7 @@ class StableDiffusionPipeline(
     ):
         super().__init__()
 
-        if hasattr(scheduler.config, "steps_offset") and scheduler.config.steps_offset != 1:
+        if scheduler is not None and getattr(scheduler.config, "steps_offset", 1) != 1:
             deprecation_message = (
                 f"The configuration file of this scheduler: {scheduler} is outdated. `steps_offset`"
                 f" should be set to 1 instead of {scheduler.config.steps_offset}. Please make sure "
@@ -223,7 +225,7 @@ class StableDiffusionPipeline(
             new_config["steps_offset"] = 1
             scheduler._internal_dict = FrozenDict(new_config)
 
-        if hasattr(scheduler.config, "clip_sample") and scheduler.config.clip_sample is True:
+        if scheduler is not None and getattr(scheduler.config, "clip_sample", False) is True:
             deprecation_message = (
                 f"The configuration file of this scheduler: {scheduler} has not set the configuration `clip_sample`."
                 " `clip_sample` should be set to False in the configuration file. Please make sure to update the"
@@ -252,12 +254,15 @@ class StableDiffusionPipeline(
                 " checker. If you do not want to use the safety checker, you can pass `'safety_checker=None'` instead."
             )
 
-        is_unet_version_less_0_9_0 = hasattr(unet.config, "_diffusers_version") and version.parse(
-            version.parse(unet.config._diffusers_version).base_version
-        ) < version.parse("0.9.0.dev0")
-        self._is_unet_config_sample_size_int = isinstance(unet.config.sample_size, int)
+        is_unet_version_less_0_9_0 = (
+            unet is not None
+            and hasattr(unet.config, "_diffusers_version")
+            and version.parse(version.parse(unet.config._diffusers_version).base_version) < version.parse("0.9.0.dev0")
+        )
+        self._is_unet_config_sample_size_int = unet is not None and isinstance(unet.config.sample_size, int)
         is_unet_sample_size_less_64 = (
-            hasattr(unet.config, "sample_size")
+            unet is not None
+            and hasattr(unet.config, "sample_size")
             and self._is_unet_config_sample_size_int
             and unet.config.sample_size < 64
         )
@@ -266,8 +271,8 @@ class StableDiffusionPipeline(
                 "The configuration file of the unet has set the default `sample_size` to smaller than"
                 " 64 which seems highly unlikely. If your checkpoint is a fine-tuned version of any of the"
                 " following: \n- CompVis/stable-diffusion-v1-4 \n- CompVis/stable-diffusion-v1-3 \n-"
-                " CompVis/stable-diffusion-v1-2 \n- CompVis/stable-diffusion-v1-1 \n- runwayml/stable-diffusion-v1-5"
-                " \n- runwayml/stable-diffusion-inpainting \n you should change 'sample_size' to 64 in the"
+                " CompVis/stable-diffusion-v1-2 \n- CompVis/stable-diffusion-v1-1 \n- stable-diffusion-v1-5/stable-diffusion-v1-5"
+                " \n- stable-diffusion-v1-5/stable-diffusion-inpainting \n you should change 'sample_size' to 64 in the"
                 " configuration file. Please make sure to update the config accordingly as leaving `sample_size=32`"
                 " in the config might lead to incorrect results in future versions. If you have downloaded this"
                 " checkpoint from the Hugging Face Hub, it would be very nice if you could open a Pull request for"
@@ -288,7 +293,7 @@ class StableDiffusionPipeline(
             feature_extractor=feature_extractor,
             image_encoder=image_encoder,
         )
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
         self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
         self.register_to_config(requires_safety_checker=requires_safety_checker)