diffusers 0.30.3__py3-none-any.whl → 0.32.0__py3-none-any.whl

diffusers/pipelines/allegro/pipeline_output.py

@@ -0,0 +1,23 @@
+from dataclasses import dataclass
+from typing import List, Union
+
+import numpy as np
+import PIL
+import torch
+
+from diffusers.utils import BaseOutput
+
+
+@dataclass
+class AllegroPipelineOutput(BaseOutput):
+    r"""
+    Output class for Allegro 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: Union[torch.Tensor, np.ndarray, List[List[PIL.Image.Image]]]

diffusers/pipelines/animatediff/__init__.py

@@ -26,6 +26,7 @@ else:
     _import_structure["pipeline_animatediff_sdxl"] = ["AnimateDiffSDXLPipeline"]
     _import_structure["pipeline_animatediff_sparsectrl"] = ["AnimateDiffSparseControlNetPipeline"]
     _import_structure["pipeline_animatediff_video2video"] = ["AnimateDiffVideoToVideoPipeline"]
+    _import_structure["pipeline_animatediff_video2video_controlnet"] = ["AnimateDiffVideoToVideoControlNetPipeline"]
 
 if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
     try:
@@ -40,6 +41,7 @@ if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
         from .pipeline_animatediff_sdxl import AnimateDiffSDXLPipeline
         from .pipeline_animatediff_sparsectrl import AnimateDiffSparseControlNetPipeline
         from .pipeline_animatediff_video2video import AnimateDiffVideoToVideoPipeline
+        from .pipeline_animatediff_video2video_controlnet import AnimateDiffVideoToVideoControlNetPipeline
         from .pipeline_output import AnimateDiffPipelineOutput
 
 else:

diffusers/pipelines/animatediff/pipeline_animatediff.py

@@ -432,7 +432,6 @@ class AnimateDiffPipeline(
             extra_step_kwargs["generator"] = generator
         return extra_step_kwargs
 
-    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.check_inputs
     def check_inputs(
         self,
         prompt,
@@ -470,8 +469,8 @@ class AnimateDiffPipeline(
             raise ValueError(
                 "Provide either `prompt` or `prompt_embeds`. Cannot leave both `prompt` and `prompt_embeds` undefined."
             )
-        elif prompt is not None and (not isinstance(prompt, str) and not isinstance(prompt, list)):
-            raise ValueError(f"`prompt` has to be of type `str` or `list` but is {type(prompt)}")
+        elif prompt is not None and not isinstance(prompt, (str, list, dict)):
+            raise ValueError(f"`prompt` has to be of type `str`, `list` or `dict` but is {type(prompt)=}")
 
         if negative_prompt is not None and negative_prompt_embeds is not None:
             raise ValueError(
@@ -557,11 +556,15 @@ class AnimateDiffPipeline(
     def num_timesteps(self):
         return self._num_timesteps
 
+    @property
+    def interrupt(self):
+        return self._interrupt
+
     @torch.no_grad()
     @replace_example_docstring(EXAMPLE_DOC_STRING)
     def __call__(
         self,
-        prompt: Union[str, List[str]] = None,
+        prompt: Optional[Union[str, List[str]]] = None,
         num_frames: Optional[int] = 16,
         height: Optional[int] = None,
         width: Optional[int] = None,
@@ -701,9 +704,10 @@ class AnimateDiffPipeline(
         self._guidance_scale = guidance_scale
         self._clip_skip = clip_skip
         self._cross_attention_kwargs = cross_attention_kwargs
+        self._interrupt = False
 
         # 2. Define call parameters
-        if prompt is not None and isinstance(prompt, str):
+        if prompt is not None and isinstance(prompt, (str, dict)):
             batch_size = 1
         elif prompt is not None and isinstance(prompt, list):
             batch_size = len(prompt)
@@ -716,22 +720,39 @@ class AnimateDiffPipeline(
         text_encoder_lora_scale = (
             self.cross_attention_kwargs.get("scale", None) if self.cross_attention_kwargs is not None else None
         )
-        prompt_embeds, negative_prompt_embeds = self.encode_prompt(
-            prompt,
-            device,
-            num_videos_per_prompt,
-            self.do_classifier_free_guidance,
-            negative_prompt,
-            prompt_embeds=prompt_embeds,
-            negative_prompt_embeds=negative_prompt_embeds,
-            lora_scale=text_encoder_lora_scale,
-            clip_skip=self.clip_skip,
-        )
-        # For classifier free guidance, we need to do two forward passes.
-        # Here we concatenate the unconditional and text embeddings into a single batch
-        # to avoid doing two forward passes
-        if self.do_classifier_free_guidance:
-            prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds])
+        if self.free_noise_enabled:
+            prompt_embeds, negative_prompt_embeds = self._encode_prompt_free_noise(
+                prompt=prompt,
+                num_frames=num_frames,
+                device=device,
+                num_videos_per_prompt=num_videos_per_prompt,
+                do_classifier_free_guidance=self.do_classifier_free_guidance,
+                negative_prompt=negative_prompt,
+                prompt_embeds=prompt_embeds,
+                negative_prompt_embeds=negative_prompt_embeds,
+                lora_scale=text_encoder_lora_scale,
+                clip_skip=self.clip_skip,
+            )
+        else:
+            prompt_embeds, negative_prompt_embeds = self.encode_prompt(
+                prompt,
+                device,
+                num_videos_per_prompt,
+                self.do_classifier_free_guidance,
+                negative_prompt,
+                prompt_embeds=prompt_embeds,
+                negative_prompt_embeds=negative_prompt_embeds,
+                lora_scale=text_encoder_lora_scale,
+                clip_skip=self.clip_skip,
+            )
+
+            # For classifier free guidance, we need to do two forward passes.
+            # Here we concatenate the unconditional and text embeddings into a single batch
+            # to avoid doing two forward passes
+            if self.do_classifier_free_guidance:
+                prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds])
+
+            prompt_embeds = prompt_embeds.repeat_interleave(repeats=num_frames, dim=0)
 
         if ip_adapter_image is not None or ip_adapter_image_embeds is not None:
             image_embeds = self.prepare_ip_adapter_image_embeds(
@@ -783,6 +804,9 @@ class AnimateDiffPipeline(
             # 8. Denoising loop
             with self.progress_bar(total=self._num_timesteps) as progress_bar:
                 for i, t in enumerate(timesteps):
+                    if self.interrupt:
+                        continue
+
                     # expand the latents if we are doing classifier free guidance
                     latent_model_input = torch.cat([latents] * 2) if self.do_classifier_free_guidance else latents
                     latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)

diffusers/pipelines/animatediff/pipeline_animatediff_controlnet.py

@@ -21,14 +21,20 @@ from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer, CLIPV
 
 from ...image_processor import PipelineImageInput
 from ...loaders import IPAdapterMixin, StableDiffusionLoraLoaderMixin, TextualInversionLoaderMixin
-from ...models import AutoencoderKL, ControlNetModel, ImageProjection, UNet2DConditionModel, UNetMotionModel
+from ...models import (
+    AutoencoderKL,
+    ControlNetModel,
+    ImageProjection,
+    MultiControlNetModel,
+    UNet2DConditionModel,
+    UNetMotionModel,
+)
 from ...models.lora import adjust_lora_scale_text_encoder
 from ...models.unets.unet_motion_model import MotionAdapter
 from ...schedulers import KarrasDiffusionSchedulers
 from ...utils import USE_PEFT_BACKEND, logging, scale_lora_layers, unscale_lora_layers
 from ...utils.torch_utils import is_compiled_module, randn_tensor
 from ...video_processor import VideoProcessor
-from ..controlnet.multicontrolnet import MultiControlNetModel
 from ..free_init_utils import FreeInitMixin
 from ..free_noise_utils import AnimateDiffFreeNoiseMixin
 from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
@@ -505,8 +511,8 @@ class AnimateDiffControlNetPipeline(
             raise ValueError(
                 "Provide either `prompt` or `prompt_embeds`. Cannot leave both `prompt` and `prompt_embeds` undefined."
             )
-        elif prompt is not None and (not isinstance(prompt, str) and not isinstance(prompt, list)):
-            raise ValueError(f"`prompt` has to be of type `str` or `list` but is {type(prompt)}")
+        elif prompt is not None and not isinstance(prompt, (str, list, dict)):
+            raise ValueError(f"`prompt` has to be of type `str`, `list` or `dict` but is {type(prompt)}")
 
         if negative_prompt is not None and negative_prompt_embeds is not None:
             raise ValueError(
@@ -699,6 +705,10 @@ class AnimateDiffControlNetPipeline(
     def num_timesteps(self):
         return self._num_timesteps
 
+    @property
+    def interrupt(self):
+        return self._interrupt
+
     @torch.no_grad()
     def __call__(
         self,
@@ -858,9 +868,10 @@ class AnimateDiffControlNetPipeline(
         self._guidance_scale = guidance_scale
         self._clip_skip = clip_skip
         self._cross_attention_kwargs = cross_attention_kwargs
+        self._interrupt = False
 
         # 2. Define call parameters
-        if prompt is not None and isinstance(prompt, str):
+        if prompt is not None and isinstance(prompt, (str, dict)):
             batch_size = 1
         elif prompt is not None and isinstance(prompt, list):
             batch_size = len(prompt)
@@ -883,22 +894,39 @@ class AnimateDiffControlNetPipeline(
         text_encoder_lora_scale = (
             cross_attention_kwargs.get("scale", None) if cross_attention_kwargs is not None else None
         )
-        prompt_embeds, negative_prompt_embeds = self.encode_prompt(
-            prompt,
-            device,
-            num_videos_per_prompt,
-            self.do_classifier_free_guidance,
-            negative_prompt,
-            prompt_embeds=prompt_embeds,
-            negative_prompt_embeds=negative_prompt_embeds,
-            lora_scale=text_encoder_lora_scale,
-            clip_skip=self.clip_skip,
-        )
-        # For classifier free guidance, we need to do two forward passes.
-        # Here we concatenate the unconditional and text embeddings into a single batch
-        # to avoid doing two forward passes
-        if self.do_classifier_free_guidance:
-            prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds])
+        if self.free_noise_enabled:
+            prompt_embeds, negative_prompt_embeds = self._encode_prompt_free_noise(
+                prompt=prompt,
+                num_frames=num_frames,
+                device=device,
+                num_videos_per_prompt=num_videos_per_prompt,
+                do_classifier_free_guidance=self.do_classifier_free_guidance,
+                negative_prompt=negative_prompt,
+                prompt_embeds=prompt_embeds,
+                negative_prompt_embeds=negative_prompt_embeds,
+                lora_scale=text_encoder_lora_scale,
+                clip_skip=self.clip_skip,
+            )
+        else:
+            prompt_embeds, negative_prompt_embeds = self.encode_prompt(
+                prompt,
+                device,
+                num_videos_per_prompt,
+                self.do_classifier_free_guidance,
+                negative_prompt,
+                prompt_embeds=prompt_embeds,
+                negative_prompt_embeds=negative_prompt_embeds,
+                lora_scale=text_encoder_lora_scale,
+                clip_skip=self.clip_skip,
+            )
+
+            # For classifier free guidance, we need to do two forward passes.
+            # Here we concatenate the unconditional and text embeddings into a single batch
+            # to avoid doing two forward passes
+            if self.do_classifier_free_guidance:
+                prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds])
+
+            prompt_embeds = prompt_embeds.repeat_interleave(repeats=num_frames, dim=0)
 
         if ip_adapter_image is not None or ip_adapter_image_embeds is not None:
             image_embeds = self.prepare_ip_adapter_image_embeds(
@@ -990,6 +1018,9 @@ class AnimateDiffControlNetPipeline(
             # 8. Denoising loop
             with self.progress_bar(total=self._num_timesteps) as progress_bar:
                 for i, t in enumerate(timesteps):
+                    if self.interrupt:
+                        continue
+
                     # expand the latents if we are doing classifier free guidance
                     latent_model_input = torch.cat([latents] * 2) if self.do_classifier_free_guidance else latents
                     latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)
@@ -1002,7 +1033,6 @@ class AnimateDiffControlNetPipeline(
                     else:
                         control_model_input = latent_model_input
                         controlnet_prompt_embeds = prompt_embeds
-                    controlnet_prompt_embeds = controlnet_prompt_embeds.repeat_interleave(num_frames, dim=0)
 
                     if isinstance(controlnet_keep[i], list):
                         cond_scale = [c * s for c, s in zip(controlnet_conditioning_scale, controlnet_keep[i])]

diffusers/pipelines/animatediff/pipeline_animatediff_sdxl.py

@@ -113,9 +113,21 @@ EXAMPLE_DOC_STRING = """
 
 # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.rescale_noise_cfg
 def rescale_noise_cfg(noise_cfg, noise_pred_text, guidance_rescale=0.0):
-    """
-    Rescale `noise_cfg` according to `guidance_rescale`. Based on findings of [Common Diffusion Noise Schedules and
-    Sample Steps are Flawed](https://arxiv.org/pdf/2305.08891.pdf). See Section 3.4
+    r"""
+    Rescales `noise_cfg` tensor based on `guidance_rescale` to improve image quality and fix overexposure. Based on
+    Section 3.4 from [Common Diffusion Noise Schedules and Sample Steps are
+    Flawed](https://arxiv.org/pdf/2305.08891.pdf).
+
+    Args:
+        noise_cfg (`torch.Tensor`):
+            The predicted noise tensor for the guided diffusion process.
+        noise_pred_text (`torch.Tensor`):
+            The predicted noise tensor for the text-guided diffusion process.
+        guidance_rescale (`float`, *optional*, defaults to 0.0):
+            A rescale factor applied to the noise predictions.
+
+    Returns:
+        noise_cfg (`torch.Tensor`): The rescaled noise prediction tensor.
     """
     std_text = noise_pred_text.std(dim=list(range(1, noise_pred_text.ndim)), keepdim=True)
     std_cfg = noise_cfg.std(dim=list(range(1, noise_cfg.ndim)), keepdim=True)
@@ -135,7 +147,7 @@ def retrieve_timesteps(
     sigmas: Optional[List[float]] = None,
     **kwargs,
 ):
-    """
+    r"""
     Calls the scheduler's `set_timesteps` method and retrieves timesteps from the scheduler after the call. Handles
     custom timesteps. Any kwargs will be supplied to `scheduler.set_timesteps`.
 
@@ -1143,6 +1155,8 @@ class AnimateDiffSDXLPipeline(
             add_text_embeds = torch.cat([negative_pooled_prompt_embeds, add_text_embeds], dim=0)
             add_time_ids = torch.cat([negative_add_time_ids, add_time_ids], dim=0)
 
+        prompt_embeds = prompt_embeds.repeat_interleave(repeats=num_frames, dim=0)
+
         prompt_embeds = prompt_embeds.to(device)
         add_text_embeds = add_text_embeds.to(device)
         add_time_ids = add_time_ids.to(device).repeat(batch_size * num_videos_per_prompt, 1)

diffusers/pipelines/animatediff/pipeline_animatediff_sparsectrl.py

@@ -24,7 +24,7 @@ from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer, CLIPV
 from ...image_processor import PipelineImageInput, VaeImageProcessor
 from ...loaders import IPAdapterMixin, StableDiffusionLoraLoaderMixin, TextualInversionLoaderMixin
 from ...models import AutoencoderKL, ImageProjection, UNet2DConditionModel, UNetMotionModel
-from ...models.controlnet_sparsectrl import SparseControlNetModel
+from ...models.controlnets.controlnet_sparsectrl import SparseControlNetModel
 from ...models.lora import adjust_lora_scale_text_encoder
 from ...models.unets.unet_motion_model import MotionAdapter
 from ...schedulers import KarrasDiffusionSchedulers
@@ -878,6 +878,8 @@ class AnimateDiffSparseControlNetPipeline(
         if self.do_classifier_free_guidance:
             prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds])
 
+        prompt_embeds = prompt_embeds.repeat_interleave(repeats=num_frames, dim=0)
+
         # 4. Prepare IP-Adapter embeddings
         if ip_adapter_image is not None or ip_adapter_image_embeds is not None:
             image_embeds = self.prepare_ip_adapter_image_embeds(

diffusers/pipelines/animatediff/pipeline_animatediff_video2video.py

@@ -119,7 +119,7 @@ def retrieve_timesteps(
     sigmas: Optional[List[float]] = None,
     **kwargs,
 ):
-    """
+    r"""
     Calls the scheduler's `set_timesteps` method and retrieves timesteps from the scheduler after the call. Handles
     custom timesteps. Any kwargs will be supplied to `scheduler.set_timesteps`.
 
@@ -246,7 +246,6 @@ class AnimateDiffVideoToVideoPipeline(
         self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
         self.video_processor = VideoProcessor(vae_scale_factor=self.vae_scale_factor)
 
-    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.encode_prompt with num_images_per_prompt -> num_videos_per_prompt
     def encode_prompt(
         self,
         prompt,
@@ -299,7 +298,7 @@ class AnimateDiffVideoToVideoPipeline(
             else:
                 scale_lora_layers(self.text_encoder, lora_scale)
 
-        if prompt is not None and isinstance(prompt, str):
+        if prompt is not None and isinstance(prompt, (str, dict)):
             batch_size = 1
         elif prompt is not None and isinstance(prompt, list):
             batch_size = len(prompt)
@@ -582,8 +581,8 @@ class AnimateDiffVideoToVideoPipeline(
             raise ValueError(
                 "Provide either `prompt` or `prompt_embeds`. Cannot leave both `prompt` and `prompt_embeds` undefined."
             )
-        elif prompt is not None and (not isinstance(prompt, str) and not isinstance(prompt, list)):
-            raise ValueError(f"`prompt` has to be of type `str` or `list` but is {type(prompt)}")
+        elif prompt is not None and not isinstance(prompt, (str, list, dict)):
+            raise ValueError(f"`prompt` has to be of type `str`, `list` or `dict` but is {type(prompt)}")
 
         if negative_prompt is not None and negative_prompt_embeds is not None:
             raise ValueError(
@@ -628,23 +627,20 @@ class AnimateDiffVideoToVideoPipeline(
 
     def prepare_latents(
         self,
-        video,
-        height,
-        width,
-        num_channels_latents,
-        batch_size,
-        timestep,
-        dtype,
-        device,
-        generator,
-        latents=None,
+        video: Optional[torch.Tensor] = None,
+        height: int = 64,
+        width: int = 64,
+        num_channels_latents: int = 4,
+        batch_size: int = 1,
+        timestep: Optional[int] = None,
+        dtype: Optional[torch.dtype] = None,
+        device: Optional[torch.device] = None,
+        generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
+        latents: Optional[torch.Tensor] = None,
         decode_chunk_size: int = 16,
-    ):
-        if latents is None:
-            num_frames = video.shape[1]
-        else:
-            num_frames = latents.shape[2]
-
+        add_noise: bool = False,
+    ) -> torch.Tensor:
+        num_frames = video.shape[1] if latents is None else latents.shape[2]
         shape = (
             batch_size,
             num_channels_latents,
@@ -666,12 +662,6 @@ class AnimateDiffVideoToVideoPipeline(
                 self.vae.to(dtype=torch.float32)
 
             if isinstance(generator, list):
-                if len(generator) != batch_size:
-                    raise ValueError(
-                        f"You have passed a list of generators of length {len(generator)}, but requested an effective batch"
-                        f" size of {batch_size}. Make sure the batch size matches the length of the generators."
-                    )
-
                 init_latents = [
                     self.encode_video(video[i], generator[i], decode_chunk_size).unsqueeze(0)
                     for i in range(batch_size)
@@ -708,8 +698,13 @@ class AnimateDiffVideoToVideoPipeline(
             if shape != latents.shape:
                 # [B, C, F, H, W]
                 raise ValueError(f"`latents` expected to have {shape=}, but found {latents.shape=}")
+
             latents = latents.to(device, dtype=dtype)
 
+            if add_noise:
+                noise = randn_tensor(shape, generator=generator, device=device, dtype=dtype)
+                latents = self.scheduler.add_noise(latents, noise, timestep)
+
         return latents
 
     @property
@@ -735,6 +730,10 @@ class AnimateDiffVideoToVideoPipeline(
     def num_timesteps(self):
         return self._num_timesteps
 
+    @property
+    def interrupt(self):
+        return self._interrupt
+
     @torch.no_grad()
     def __call__(
         self,
@@ -743,6 +742,7 @@ class AnimateDiffVideoToVideoPipeline(
         height: Optional[int] = None,
         width: Optional[int] = None,
         num_inference_steps: int = 50,
+        enforce_inference_steps: bool = False,
         timesteps: Optional[List[int]] = None,
         sigmas: Optional[List[float]] = None,
         guidance_scale: float = 7.5,
@@ -874,9 +874,10 @@ class AnimateDiffVideoToVideoPipeline(
         self._guidance_scale = guidance_scale
         self._clip_skip = clip_skip
         self._cross_attention_kwargs = cross_attention_kwargs
+        self._interrupt = False
 
         # 2. Define call parameters
-        if prompt is not None and isinstance(prompt, str):
+        if prompt is not None and isinstance(prompt, (str, dict)):
             batch_size = 1
         elif prompt is not None and isinstance(prompt, list):
             batch_size = len(prompt)
@@ -884,51 +885,29 @@ class AnimateDiffVideoToVideoPipeline(
             batch_size = prompt_embeds.shape[0]
 
         device = self._execution_device
+        dtype = self.dtype
 
-        # 3. Encode input prompt
-        text_encoder_lora_scale = (
-            self.cross_attention_kwargs.get("scale", None) if self.cross_attention_kwargs is not None else None
-        )
-        prompt_embeds, negative_prompt_embeds = self.encode_prompt(
-            prompt,
-            device,
-            num_videos_per_prompt,
-            self.do_classifier_free_guidance,
-            negative_prompt,
-            prompt_embeds=prompt_embeds,
-            negative_prompt_embeds=negative_prompt_embeds,
-            lora_scale=text_encoder_lora_scale,
-            clip_skip=self.clip_skip,
-        )
-
-        # For classifier free guidance, we need to do two forward passes.
-        # Here we concatenate the unconditional and text embeddings into a single batch
-        # to avoid doing two forward passes
-        if self.do_classifier_free_guidance:
-            prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds])
-
-        if ip_adapter_image is not None or ip_adapter_image_embeds is not None:
-            image_embeds = self.prepare_ip_adapter_image_embeds(
-                ip_adapter_image,
-                ip_adapter_image_embeds,
-                device,
-                batch_size * num_videos_per_prompt,
-                self.do_classifier_free_guidance,
+        # 3. Prepare timesteps
+        if not enforce_inference_steps:
+            timesteps, num_inference_steps = retrieve_timesteps(
+                self.scheduler, num_inference_steps, device, timesteps, sigmas
             )
+            timesteps, num_inference_steps = self.get_timesteps(num_inference_steps, timesteps, strength, device)
+            latent_timestep = timesteps[:1].repeat(batch_size * num_videos_per_prompt)
+        else:
+            denoising_inference_steps = int(num_inference_steps / strength)
+            timesteps, denoising_inference_steps = retrieve_timesteps(
+                self.scheduler, denoising_inference_steps, device, timesteps, sigmas
+            )
+            timesteps = timesteps[-num_inference_steps:]
+            latent_timestep = timesteps[:1].repeat(batch_size * num_videos_per_prompt)
 
-        # 4. Prepare timesteps
-        timesteps, num_inference_steps = retrieve_timesteps(
-            self.scheduler, num_inference_steps, device, timesteps, sigmas
-        )
-        timesteps, num_inference_steps = self.get_timesteps(num_inference_steps, timesteps, strength, device)
-        latent_timestep = timesteps[:1].repeat(batch_size * num_videos_per_prompt)
-
-        # 5. Prepare latent variables
+        # 4. Prepare latent variables
         if latents is None:
             video = self.video_processor.preprocess_video(video, height=height, width=width)
             # Move the number of frames before the number of channels.
             video = video.permute(0, 2, 1, 3, 4)
-            video = video.to(device=device, dtype=prompt_embeds.dtype)
+            video = video.to(device=device, dtype=dtype)
         num_channels_latents = self.unet.config.in_channels
         latents = self.prepare_latents(
             video=video,
@@ -937,17 +916,67 @@ class AnimateDiffVideoToVideoPipeline(
             num_channels_latents=num_channels_latents,
             batch_size=batch_size * num_videos_per_prompt,
             timestep=latent_timestep,
-            dtype=prompt_embeds.dtype,
+            dtype=dtype,
             device=device,
             generator=generator,
             latents=latents,
             decode_chunk_size=decode_chunk_size,
+            add_noise=enforce_inference_steps,
         )
 
-        # 6. Prepare extra step kwargs. TODO: Logic should ideally just be moved out of the pipeline
+        # 5. Encode input prompt
+        text_encoder_lora_scale = (
+            self.cross_attention_kwargs.get("scale", None) if self.cross_attention_kwargs is not None else None
+        )
+        num_frames = latents.shape[2]
+        if self.free_noise_enabled:
+            prompt_embeds, negative_prompt_embeds = self._encode_prompt_free_noise(
+                prompt=prompt,
+                num_frames=num_frames,
+                device=device,
+                num_videos_per_prompt=num_videos_per_prompt,
+                do_classifier_free_guidance=self.do_classifier_free_guidance,
+                negative_prompt=negative_prompt,
+                prompt_embeds=prompt_embeds,
+                negative_prompt_embeds=negative_prompt_embeds,
+                lora_scale=text_encoder_lora_scale,
+                clip_skip=self.clip_skip,
+            )
+        else:
+            prompt_embeds, negative_prompt_embeds = self.encode_prompt(
+                prompt,
+                device,
+                num_videos_per_prompt,
+                self.do_classifier_free_guidance,
+                negative_prompt,
+                prompt_embeds=prompt_embeds,
+                negative_prompt_embeds=negative_prompt_embeds,
+                lora_scale=text_encoder_lora_scale,
+                clip_skip=self.clip_skip,
+            )
+
+            # For classifier free guidance, we need to do two forward passes.
+            # Here we concatenate the unconditional and text embeddings into a single batch
+            # to avoid doing two forward passes
+            if self.do_classifier_free_guidance:
+                prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds])
+
+            prompt_embeds = prompt_embeds.repeat_interleave(repeats=num_frames, dim=0)
+
+        # 6. Prepare IP-Adapter embeddings
+        if ip_adapter_image is not None or ip_adapter_image_embeds is not None:
+            image_embeds = self.prepare_ip_adapter_image_embeds(
+                ip_adapter_image,
+                ip_adapter_image_embeds,
+                device,
+                batch_size * num_videos_per_prompt,
+                self.do_classifier_free_guidance,
+            )
+
+        # 7. Prepare extra step kwargs. TODO: Logic should ideally just be moved out of the pipeline
         extra_step_kwargs = self.prepare_extra_step_kwargs(generator, eta)
 
-        # 7. Add image embeds for IP-Adapter
+        # 8. Add image embeds for IP-Adapter
         added_cond_kwargs = (
             {"image_embeds": image_embeds}
             if ip_adapter_image is not None or ip_adapter_image_embeds is not None
@@ -967,9 +996,12 @@ class AnimateDiffVideoToVideoPipeline(
             self._num_timesteps = len(timesteps)
             num_warmup_steps = len(timesteps) - num_inference_steps * self.scheduler.order
 
-            # 8. Denoising loop
+            # 9. Denoising loop
             with self.progress_bar(total=self._num_timesteps) as progress_bar:
                 for i, t in enumerate(timesteps):
+                    if self.interrupt:
+                        continue
+
                     # expand the latents if we are doing classifier free guidance
                     latent_model_input = torch.cat([latents] * 2) if self.do_classifier_free_guidance else latents
                     latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)
@@ -1005,14 +1037,14 @@ class AnimateDiffVideoToVideoPipeline(
                     if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0):
                         progress_bar.update()
 
-        # 9. Post-processing
+        # 10. Post-processing
         if output_type == "latent":
             video = latents
         else:
             video_tensor = self.decode_latents(latents, decode_chunk_size)
             video = self.video_processor.postprocess_video(video=video_tensor, output_type=output_type)
 
-        # 10. Offload all models
+        # 11. Offload all models
         self.maybe_free_model_hooks()
 
         if not return_dict: