diffusers 0.31.0__py3-none-any.whl → 0.32.1__py3-none-any.whl

diffusers/pipelines/mochi/pipeline_output.py

@@ -0,0 +1,20 @@
+from dataclasses import dataclass
+
+import torch
+
+from diffusers.utils import BaseOutput
+
+
+@dataclass
+class MochiPipelineOutput(BaseOutput):
+    r"""
+    Output class for Mochi pipelines.
+
+    Args:
+        frames (`torch.Tensor`, `np.ndarray`, or List[List[PIL.Image.Image]]):
+            List of video outputs - It can be a nested list of length `batch_size,` with each sub-list containing
+            denoised PIL image sequences of length `num_frames.` It can also be a NumPy array or Torch tensor of shape
+            `(batch_size, num_frames, channels, height, width)`.
+    """
+
+    frames: torch.Tensor

diffusers/pipelines/pag/__init__.py

@@ -29,10 +29,14 @@ else:
     _import_structure["pipeline_pag_hunyuandit"] = ["HunyuanDiTPAGPipeline"]
     _import_structure["pipeline_pag_kolors"] = ["KolorsPAGPipeline"]
     _import_structure["pipeline_pag_pixart_sigma"] = ["PixArtSigmaPAGPipeline"]
+    _import_structure["pipeline_pag_sana"] = ["SanaPAGPipeline"]
     _import_structure["pipeline_pag_sd"] = ["StableDiffusionPAGPipeline"]
     _import_structure["pipeline_pag_sd_3"] = ["StableDiffusion3PAGPipeline"]
+    _import_structure["pipeline_pag_sd_3_img2img"] = ["StableDiffusion3PAGImg2ImgPipeline"]
     _import_structure["pipeline_pag_sd_animatediff"] = ["AnimateDiffPAGPipeline"]
     _import_structure["pipeline_pag_sd_img2img"] = ["StableDiffusionPAGImg2ImgPipeline"]
+    _import_structure["pipeline_pag_sd_inpaint"] = ["StableDiffusionPAGInpaintPipeline"]
+
     _import_structure["pipeline_pag_sd_xl"] = ["StableDiffusionXLPAGPipeline"]
     _import_structure["pipeline_pag_sd_xl_img2img"] = ["StableDiffusionXLPAGImg2ImgPipeline"]
     _import_structure["pipeline_pag_sd_xl_inpaint"] = ["StableDiffusionXLPAGInpaintPipeline"]
@@ -52,10 +56,13 @@ if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
         from .pipeline_pag_hunyuandit import HunyuanDiTPAGPipeline
         from .pipeline_pag_kolors import KolorsPAGPipeline
         from .pipeline_pag_pixart_sigma import PixArtSigmaPAGPipeline
+        from .pipeline_pag_sana import SanaPAGPipeline
         from .pipeline_pag_sd import StableDiffusionPAGPipeline
         from .pipeline_pag_sd_3 import StableDiffusion3PAGPipeline
+        from .pipeline_pag_sd_3_img2img import StableDiffusion3PAGImg2ImgPipeline
         from .pipeline_pag_sd_animatediff import AnimateDiffPAGPipeline
         from .pipeline_pag_sd_img2img import StableDiffusionPAGImg2ImgPipeline
+        from .pipeline_pag_sd_inpaint import StableDiffusionPAGInpaintPipeline
         from .pipeline_pag_sd_xl import StableDiffusionXLPAGPipeline
         from .pipeline_pag_sd_xl_img2img import StableDiffusionXLPAGImg2ImgPipeline
         from .pipeline_pag_sd_xl_inpaint import StableDiffusionXLPAGInpaintPipeline

diffusers/pipelines/pag/pipeline_pag_controlnet_sd.py

@@ -25,7 +25,7 @@ from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer, CLIPV
 from ...callbacks import MultiPipelineCallbacks, PipelineCallback
 from ...image_processor import PipelineImageInput, VaeImageProcessor
 from ...loaders import FromSingleFileMixin, IPAdapterMixin, StableDiffusionLoraLoaderMixin, TextualInversionLoaderMixin
-from ...models import AutoencoderKL, ControlNetModel, ImageProjection, UNet2DConditionModel
+from ...models import AutoencoderKL, ControlNetModel, ImageProjection, MultiControlNetModel, UNet2DConditionModel
 from ...models.lora import adjust_lora_scale_text_encoder
 from ...schedulers import KarrasDiffusionSchedulers
 from ...utils import (
@@ -36,7 +36,6 @@ from ...utils import (
     unscale_lora_layers,
 )
 from ...utils.torch_utils import is_compiled_module, is_torch_version, randn_tensor
-from ..controlnet.multicontrolnet import MultiControlNetModel
 from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
 from ..stable_diffusion.pipeline_output import StableDiffusionPipelineOutput
 from ..stable_diffusion.safety_checker import StableDiffusionSafetyChecker

diffusers/pipelines/pag/pipeline_pag_controlnet_sd_inpaint.py

@@ -26,7 +26,7 @@ from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer, CLIPV
 from ...callbacks import MultiPipelineCallbacks, PipelineCallback
 from ...image_processor import PipelineImageInput, VaeImageProcessor
 from ...loaders import FromSingleFileMixin, IPAdapterMixin, StableDiffusionLoraLoaderMixin, TextualInversionLoaderMixin
-from ...models import AutoencoderKL, ControlNetModel, ImageProjection, UNet2DConditionModel
+from ...models import AutoencoderKL, ControlNetModel, ImageProjection, MultiControlNetModel, UNet2DConditionModel
 from ...models.lora import adjust_lora_scale_text_encoder
 from ...schedulers import KarrasDiffusionSchedulers
 from ...utils import (
@@ -37,7 +37,6 @@ from ...utils import (
     unscale_lora_layers,
 )
 from ...utils.torch_utils import is_compiled_module, randn_tensor
-from ..controlnet.multicontrolnet import MultiControlNetModel
 from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
 from ..stable_diffusion import StableDiffusionPipelineOutput
 from ..stable_diffusion.safety_checker import StableDiffusionSafetyChecker

diffusers/pipelines/pag/pipeline_pag_controlnet_sd_xl.py

@@ -38,7 +38,7 @@ from ...loaders import (
     StableDiffusionXLLoraLoaderMixin,
     TextualInversionLoaderMixin,
 )
-from ...models import AutoencoderKL, ControlNetModel, ImageProjection, UNet2DConditionModel
+from ...models import AutoencoderKL, ControlNetModel, ImageProjection, MultiControlNetModel, UNet2DConditionModel
 from ...models.attention_processor import (
     AttnProcessor2_0,
     XFormersAttnProcessor,
@@ -61,8 +61,6 @@ from .pag_utils import PAGMixin
 if is_invisible_watermark_available():
     from ..stable_diffusion_xl.watermark import StableDiffusionXLWatermarker
 
-from ..controlnet.multicontrolnet import MultiControlNetModel
-
 
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 

diffusers/pipelines/pag/pipeline_pag_controlnet_sd_xl_img2img.py

@@ -38,7 +38,7 @@ from ...loaders import (
     StableDiffusionXLLoraLoaderMixin,
     TextualInversionLoaderMixin,
 )
-from ...models import AutoencoderKL, ControlNetModel, ImageProjection, UNet2DConditionModel
+from ...models import AutoencoderKL, ControlNetModel, ImageProjection, MultiControlNetModel, UNet2DConditionModel
 from ...models.attention_processor import (
     AttnProcessor2_0,
     XFormersAttnProcessor,
@@ -61,8 +61,6 @@ from .pag_utils import PAGMixin
 if is_invisible_watermark_available():
     from ..stable_diffusion_xl.watermark import StableDiffusionXLWatermarker
 
-from ..controlnet.multicontrolnet import MultiControlNetModel
-
 
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 

diffusers/pipelines/pag/pipeline_pag_hunyuandit.py

@@ -818,7 +818,11 @@ class HunyuanDiTPAGPipeline(DiffusionPipeline, PAGMixin):
         base_size = 512 // 8 // self.transformer.config.patch_size
         grid_crops_coords = get_resize_crop_region_for_grid((grid_height, grid_width), base_size)
         image_rotary_emb = get_2d_rotary_pos_embed(
-            self.transformer.inner_dim // self.transformer.num_heads, grid_crops_coords, (grid_height, grid_width)
+            self.transformer.inner_dim // self.transformer.num_heads,
+            grid_crops_coords,
+            (grid_height, grid_width),
+            device=device,
+            output_type="pt",
         )
 
         style = torch.tensor([0], device=device)

diffusers/pipelines/pag/pipeline_pag_pixart_sigma.py

@@ -227,13 +227,6 @@ class PixArtSigmaPAGPipeline(DiffusionPipeline, PAGMixin):
         if device is None:
             device = self._execution_device
 
-        if prompt is not None and isinstance(prompt, str):
-            batch_size = 1
-        elif prompt is not None and isinstance(prompt, list):
-            batch_size = len(prompt)
-        else:
-            batch_size = prompt_embeds.shape[0]
-
         # See Section 3.1. of the paper.
         max_length = max_sequence_length
 
@@ -278,12 +271,12 @@ class PixArtSigmaPAGPipeline(DiffusionPipeline, PAGMixin):
         # duplicate text embeddings and attention mask for each generation per prompt, using mps friendly method
         prompt_embeds = prompt_embeds.repeat(1, num_images_per_prompt, 1)
         prompt_embeds = prompt_embeds.view(bs_embed * num_images_per_prompt, seq_len, -1)
-        prompt_attention_mask = prompt_attention_mask.view(bs_embed, -1)
-        prompt_attention_mask = prompt_attention_mask.repeat(num_images_per_prompt, 1)
+        prompt_attention_mask = prompt_attention_mask.repeat(1, num_images_per_prompt)
+        prompt_attention_mask = prompt_attention_mask.view(bs_embed * num_images_per_prompt, -1)
 
         # get unconditional embeddings for classifier free guidance
         if do_classifier_free_guidance and negative_prompt_embeds is None:
-            uncond_tokens = [negative_prompt] * batch_size if isinstance(negative_prompt, str) else negative_prompt
+            uncond_tokens = [negative_prompt] * bs_embed if isinstance(negative_prompt, str) else negative_prompt
             uncond_tokens = self._text_preprocessing(uncond_tokens, clean_caption=clean_caption)
             max_length = prompt_embeds.shape[1]
             uncond_input = self.tokenizer(
@@ -310,10 +303,10 @@ class PixArtSigmaPAGPipeline(DiffusionPipeline, PAGMixin):
             negative_prompt_embeds = negative_prompt_embeds.to(dtype=dtype, device=device)
 
             negative_prompt_embeds = negative_prompt_embeds.repeat(1, num_images_per_prompt, 1)
-            negative_prompt_embeds = negative_prompt_embeds.view(batch_size * num_images_per_prompt, seq_len, -1)
+            negative_prompt_embeds = negative_prompt_embeds.view(bs_embed * num_images_per_prompt, seq_len, -1)
 
-            negative_prompt_attention_mask = negative_prompt_attention_mask.view(bs_embed, -1)
-            negative_prompt_attention_mask = negative_prompt_attention_mask.repeat(num_images_per_prompt, 1)
+            negative_prompt_attention_mask = negative_prompt_attention_mask.repeat(1, num_images_per_prompt)
+            negative_prompt_attention_mask = negative_prompt_attention_mask.view(bs_embed * num_images_per_prompt, -1)
         else:
             negative_prompt_embeds = None
             negative_prompt_attention_mask = None