Merge branch 'main' into integrations/first-block-cache-2

Author: a-r-r-o-w

docs/source/en/api/pipelines/wan.md

@@ -133,6 +133,46 @@ output = pipe(
 export_to_video(output, "wan-i2v.mp4", fps=16)
 ```
 
+### Video to Video Generation
+
+```python
+import torch
+from diffusers.utils import load_video, export_to_video
+from diffusers import AutoencoderKLWan, WanVideoToVideoPipeline, UniPCMultistepScheduler
+
+# Available models: Wan-AI/Wan2.1-T2V-14B-Diffusers, Wan-AI/Wan2.1-T2V-1.3B-Diffusers
+model_id = "Wan-AI/Wan2.1-T2V-1.3B-Diffusers"
+vae = AutoencoderKLWan.from_pretrained(
+    model_id, subfolder="vae", torch_dtype=torch.float32
+)
+pipe = WanVideoToVideoPipeline.from_pretrained(
+    model_id, vae=vae, torch_dtype=torch.bfloat16
+)
+flow_shift = 3.0  # 5.0 for 720P, 3.0 for 480P
+pipe.scheduler = UniPCMultistepScheduler.from_config(
+    pipe.scheduler.config, flow_shift=flow_shift
+)
+# change to pipe.to("cuda") if you have sufficient VRAM
+pipe.enable_model_cpu_offload()
+
+prompt = "A robot standing on a mountain top. The sun is setting in the background"
+negative_prompt = "Bright tones, overexposed, static, blurred details, subtitles, style, works, paintings, images, static, overall gray, worst quality, low quality, JPEG compression residue, ugly, incomplete, extra fingers, poorly drawn hands, poorly drawn faces, deformed, disfigured, misshapen limbs, fused fingers, still picture, messy background, three legs, many people in the background, walking backwards"
+video = load_video(
+    "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/hiker.mp4"
+)
+output = pipe(
+    video=video,
+    prompt=prompt,
+    negative_prompt=negative_prompt,
+    height=480,
+    width=512,
+    guidance_scale=7.0,
+    strength=0.7,
+).frames[0]
+
+export_to_video(output, "wan-v2v.mp4", fps=16)
+```
+
 ## Memory Optimizations for Wan 2.1
 
 Base inference with the large 14B Wan 2.1 models can take up to 35GB of VRAM when generating videos at 720p resolution. We'll outline a few memory optimizations we can apply to reduce the VRAM required to run the model.
@@ -323,7 +363,7 @@ import numpy as np
 from diffusers import AutoencoderKLWan, WanTransformer3DModel, WanImageToVideoPipeline
 from diffusers.hooks.group_offloading import apply_group_offloading
 from diffusers.utils import export_to_video, load_image
-from transformers import UMT5EncoderModel, CLIPVisionMode
+from transformers import UMT5EncoderModel, CLIPVisionModel
 
 model_id = "Wan-AI/Wan2.1-I2V-14B-720P-Diffusers"
 image_encoder = CLIPVisionModel.from_pretrained(
@@ -356,7 +396,7 @@ prompt = (
     "An astronaut hatching from an egg, on the surface of the moon, the darkness and depth of space realised in "
     "the background. High quality, ultrarealistic detail and breath-taking movie-like camera shot."
 )
-negative_prompt = "Bright tones, overexposed, static, blurred details, subtitles, style, works, paintings, images, static, overall gray, worst quality, low quality, JPEG compression residue, ugly, incomplete, extra fingers, poorly drawn hands, poorly drawn faces, deformed, disfigured, misshapen limbs, fused fingers, still picture, messy background, three legs, many people in the background, walking backwards
+negative_prompt = "Bright tones, overexposed, static, blurred details, subtitles, style, works, paintings, images, static, overall gray, worst quality, low quality, JPEG compression residue, ugly, incomplete, extra fingers, poorly drawn hands, poorly drawn faces, deformed, disfigured, misshapen limbs, fused fingers, still picture, messy background, three legs, many people in the background, walking backwards"
 num_frames = 33
 
 output = pipe(
@@ -372,7 +412,7 @@ output = pipe(
 export_to_video(output, "wan-i2v.mp4", fps=16)
 ```
 
-### Using a Custom Scheduler
+## Using a Custom Scheduler
 
 Wan can be used with many different schedulers, each with their own benefits regarding speed and generation quality. By default, Wan uses the `UniPCMultistepScheduler(prediction_type="flow_prediction", use_flow_sigmas=True, flow_shift=3.0)` scheduler. You can use a different scheduler as follows:
 
@@ -403,7 +443,7 @@ transformer = WanTransformer3DModel.from_single_file(ckpt_path, torch_dtype=torc
 pipe = WanPipeline.from_pretrained("Wan-AI/Wan2.1-T2V-1.3B-Diffusers", transformer=transformer)
 ```
 
-## Recommendations for Inference:
+## Recommendations for Inference
 - Keep `AutencoderKLWan` in `torch.float32` for better decoding quality.
 - `num_frames` should satisfy the following constraint: `(num_frames - 1) % 4 == 0`
 - For smaller resolution videos, try lower values of `shift` (between `2.0` to `5.0`) in the [Scheduler](https://huggingface.co/docs/diffusers/main/en/api/schedulers/flow_match_euler_discrete#diffusers.FlowMatchEulerDiscreteScheduler.shift). For larger resolution videos, try higher values (between `7.0` and `12.0`). The default value is `3.0` for Wan.

docs/source/en/installation.md

@@ -161,10 +161,10 @@ Your Python environment will find the `main` version of 🤗 Diffusers on the ne
 
 Model weights and files are downloaded from the Hub to a cache which is usually your home directory. You can change the cache location by specifying the `HF_HOME` or `HUGGINFACE_HUB_CACHE` environment variables or configuring the `cache_dir` parameter in methods like [`~DiffusionPipeline.from_pretrained`].
 
-Cached files allow you to run 🤗 Diffusers offline. To prevent 🤗 Diffusers from connecting to the internet, set the `HF_HUB_OFFLINE` environment variable to `True` and 🤗 Diffusers will only load previously downloaded files in the cache.
+Cached files allow you to run 🤗 Diffusers offline. To prevent 🤗 Diffusers from connecting to the internet, set the `HF_HUB_OFFLINE` environment variable to `1` and 🤗 Diffusers will only load previously downloaded files in the cache.
 
 ```shell
-export HF_HUB_OFFLINE=True
+export HF_HUB_OFFLINE=1
 ```
 
 For more details about managing and cleaning the cache, take a look at the [caching](https://huggingface.co/docs/huggingface_hub/guides/manage-cache) guide.
@@ -179,14 +179,16 @@ Telemetry is only sent when loading models and pipelines from the Hub,
 and it is not collected if you're loading local files.
 
 We understand that not everyone wants to share additional information,and we respect your privacy.
-You can disable telemetry collection by setting the `DISABLE_TELEMETRY` environment variable from your terminal:
+You can disable telemetry collection by setting the `HF_HUB_DISABLE_TELEMETRY` environment variable from your terminal:
 
 On Linux/MacOS:
+
 ```bash
-export DISABLE_TELEMETRY=YES
+export HF_HUB_DISABLE_TELEMETRY=1
 ```
 
 On Windows:
+
 ```bash
-set DISABLE_TELEMETRY=YES
+set HF_HUB_DISABLE_TELEMETRY=1
 ```

examples/community/lpw_stable_diffusion_xl.py

@@ -1773,7 +1773,7 @@ def denoising_value_valid(dnv):
                         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."
                     )
             elif num_channels_unet != 4:
@@ -1924,7 +1924,22 @@ def denoising_value_valid(dnv):
                 self.upcast_vae()
                 latents = latents.to(next(iter(self.vae.post_quant_conv.parameters())).dtype)
 
-            image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False)[0]
+            # unscale/denormalize the latents
+            # denormalize with the mean and std if available and not None
+            has_latents_mean = hasattr(self.vae.config, "latents_mean") and self.vae.config.latents_mean is not None
+            has_latents_std = hasattr(self.vae.config, "latents_std") and self.vae.config.latents_std is not None
+            if has_latents_mean and has_latents_std:
+                latents_mean = (
+                    torch.tensor(self.vae.config.latents_mean).view(1, 4, 1, 1).to(latents.device, latents.dtype)
+                )
+                latents_std = (
+                    torch.tensor(self.vae.config.latents_std).view(1, 4, 1, 1).to(latents.device, latents.dtype)
+                )
+                latents = latents * latents_std / self.vae.config.scaling_factor + latents_mean
+            else:
+                latents = latents / self.vae.config.scaling_factor
+
+            image = self.vae.decode(latents, return_dict=False)[0]
 
             # cast back to fp16 if needed
             if needs_upcasting:

src/diffusers/__init__.py

@@ -511,6 +511,7 @@
             "VQDiffusionPipeline",
             "WanImageToVideoPipeline",
             "WanPipeline",
+            "WanVideoToVideoPipeline",
             "WuerstchenCombinedPipeline",
             "WuerstchenDecoderPipeline",
             "WuerstchenPriorPipeline",
@@ -1066,6 +1067,7 @@
             VQDiffusionPipeline,
             WanImageToVideoPipeline,
             WanPipeline,
+            WanVideoToVideoPipeline,
             WuerstchenCombinedPipeline,
             WuerstchenDecoderPipeline,
             WuerstchenPriorPipeline,

src/diffusers/pipelines/__init__.py

@@ -356,7 +356,7 @@
         "WuerstchenDecoderPipeline",
         "WuerstchenPriorPipeline",
     ]
-    _import_structure["wan"] = ["WanPipeline", "WanImageToVideoPipeline"]
+    _import_structure["wan"] = ["WanPipeline", "WanImageToVideoPipeline", "WanVideoToVideoPipeline"]
 try:
     if not is_onnx_available():
         raise OptionalDependencyNotAvailable()
@@ -709,7 +709,7 @@
             UniDiffuserPipeline,
             UniDiffuserTextDecoder,
         )
-        from .wan import WanImageToVideoPipeline, WanPipeline
+        from .wan import WanImageToVideoPipeline, WanPipeline, WanVideoToVideoPipeline
         from .wuerstchen import (
             WuerstchenCombinedPipeline,
             WuerstchenDecoderPipeline,

src/diffusers/pipelines/ltx/pipeline_ltx_image2video.py

@@ -487,19 +487,21 @@ def prepare_latents(
     ) -> torch.Tensor:
         height = height // self.vae_spatial_compression_ratio
         width = width // self.vae_spatial_compression_ratio
-        num_frames = (
-            (num_frames - 1) // self.vae_temporal_compression_ratio + 1 if latents is None else latents.size(2)
-        )
+        num_frames = (num_frames - 1) // self.vae_temporal_compression_ratio + 1
 
         shape = (batch_size, num_channels_latents, num_frames, height, width)
         mask_shape = (batch_size, 1, num_frames, height, width)
 
         if latents is not None:
-            conditioning_mask = latents.new_zeros(shape)
+            conditioning_mask = latents.new_zeros(mask_shape)
             conditioning_mask[:, :, 0] = 1.0
             conditioning_mask = self._pack_latents(
                 conditioning_mask, self.transformer_spatial_patch_size, self.transformer_temporal_patch_size
-            )
+            ).squeeze(-1)
+            if latents.ndim != 3 or latents.shape[:2] != conditioning_mask.shape:
+                raise ValueError(
+                    f"Provided `latents` tensor has shape {latents.shape}, but the expected shape is {conditioning_mask.shape + (num_channels_latents,)}."
+                )
             return latents.to(device=device, dtype=dtype), conditioning_mask
 
         if isinstance(generator, list):

src/diffusers/pipelines/wan/__init__.py

@@ -24,7 +24,7 @@
 else:
     _import_structure["pipeline_wan"] = ["WanPipeline"]
     _import_structure["pipeline_wan_i2v"] = ["WanImageToVideoPipeline"]
-
+    _import_structure["pipeline_wan_video2video"] = ["WanVideoToVideoPipeline"]
 if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
     try:
         if not (is_transformers_available() and is_torch_available()):
@@ -35,6 +35,7 @@
     else:
         from .pipeline_wan import WanPipeline
         from .pipeline_wan_i2v import WanImageToVideoPipeline
+        from .pipeline_wan_video2video import WanVideoToVideoPipeline
 
 else:
     import sys

src/diffusers/pipelines/wan/pipeline_wan.py

@@ -458,6 +458,13 @@ def __call__(
             callback_on_step_end_tensor_inputs,
         )
 
+        if num_frames % self.vae_scale_factor_temporal != 1:
+            logger.warning(
+                f"`num_frames - 1` has to be divisible by {self.vae_scale_factor_temporal}. Rounding to the nearest number."
+            )
+            num_frames = num_frames // self.vae_scale_factor_temporal * self.vae_scale_factor_temporal + 1
+        num_frames = max(num_frames, 1)
+
         self._guidance_scale = guidance_scale
         self._attention_kwargs = attention_kwargs
         self._current_timestep = None

src/diffusers/pipelines/wan/pipeline_wan_i2v.py

@@ -559,6 +559,13 @@ def __call__(
             callback_on_step_end_tensor_inputs,
         )
 
+        if num_frames % self.vae_scale_factor_temporal != 1:
+            logger.warning(
+                f"`num_frames - 1` has to be divisible by {self.vae_scale_factor_temporal}. Rounding to the nearest number."
+            )
+            num_frames = num_frames // self.vae_scale_factor_temporal * self.vae_scale_factor_temporal + 1
+        num_frames = max(num_frames, 1)
+
         self._guidance_scale = guidance_scale
         self._attention_kwargs = attention_kwargs
         self._current_timestep = None