Merge branch 'main' of https://github.com/huggingface/diffusers into main

Author: Skquark

docs/source/en/_toctree.yml

@@ -290,12 +290,12 @@
         title: AuraFlowTransformer2DModel
       - local: api/models/cogvideox_transformer3d
         title: CogVideoXTransformer3DModel
-      - local: api/models/consisid_transformer3d
-        title: ConsisIDTransformer3DModel
       - local: api/models/cogview3plus_transformer2d
         title: CogView3PlusTransformer2DModel
       - local: api/models/cogview4_transformer2d
         title: CogView4Transformer2DModel
+      - local: api/models/consisid_transformer3d
+        title: ConsisIDTransformer3DModel
       - local: api/models/dit_transformer2d
         title: DiTTransformer2DModel
       - local: api/models/easyanimate_transformer3d
@@ -310,12 +310,12 @@
         title: HunyuanVideoTransformer3DModel
       - local: api/models/latte_transformer3d
         title: LatteTransformer3DModel
-      - local: api/models/lumina_nextdit2d
-        title: LuminaNextDiT2DModel
-      - local: api/models/lumina2_transformer2d
-        title: Lumina2Transformer2DModel
       - local: api/models/ltx_video_transformer3d
         title: LTXVideoTransformer3DModel
+      - local: api/models/lumina2_transformer2d
+        title: Lumina2Transformer2DModel
+      - local: api/models/lumina_nextdit2d
+        title: LuminaNextDiT2DModel
       - local: api/models/mochi_transformer3d
         title: MochiTransformer3DModel
       - local: api/models/omnigen_transformer
@@ -324,10 +324,10 @@
         title: PixArtTransformer2DModel
       - local: api/models/prior_transformer
         title: PriorTransformer
-      - local: api/models/sd3_transformer2d
-        title: SD3Transformer2DModel
       - local: api/models/sana_transformer2d
         title: SanaTransformer2DModel
+      - local: api/models/sd3_transformer2d
+        title: SD3Transformer2DModel
       - local: api/models/stable_audio_transformer
         title: StableAudioDiTModel
       - local: api/models/transformer2d
@@ -342,10 +342,10 @@
         title: StableCascadeUNet
       - local: api/models/unet
         title: UNet1DModel
-      - local: api/models/unet2d
-        title: UNet2DModel
       - local: api/models/unet2d-cond
         title: UNet2DConditionModel
+      - local: api/models/unet2d
+        title: UNet2DModel
       - local: api/models/unet3d-cond
         title: UNet3DConditionModel
       - local: api/models/unet-motion
@@ -354,6 +354,10 @@
         title: UViT2DModel
       title: UNets
     - sections:
+      - local: api/models/asymmetricautoencoderkl
+        title: AsymmetricAutoencoderKL
+      - local: api/models/autoencoder_dc
+        title: AutoencoderDC
       - local: api/models/autoencoderkl
         title: AutoencoderKL
       - local: api/models/autoencoderkl_allegro
@@ -370,10 +374,6 @@
         title: AutoencoderKLMochi
       - local: api/models/autoencoder_kl_wan
         title: AutoencoderKLWan
-      - local: api/models/asymmetricautoencoderkl
-        title: AsymmetricAutoencoderKL
-      - local: api/models/autoencoder_dc
-        title: AutoencoderDC
       - local: api/models/consistency_decoder_vae
         title: ConsistencyDecoderVAE
       - local: api/models/autoencoder_oobleck
@@ -521,40 +521,40 @@
     - sections:
       - local: api/pipelines/stable_diffusion/overview
         title: Overview
-      - local: api/pipelines/stable_diffusion/text2img
-        title: Text-to-image
+      - local: api/pipelines/stable_diffusion/depth2img
+        title: Depth-to-image
+      - local: api/pipelines/stable_diffusion/gligen
+        title: GLIGEN (Grounded Language-to-Image Generation)
+      - local: api/pipelines/stable_diffusion/image_variation
+        title: Image variation
       - local: api/pipelines/stable_diffusion/img2img
         title: Image-to-image
       - local: api/pipelines/stable_diffusion/svd
         title: Image-to-video
       - local: api/pipelines/stable_diffusion/inpaint
         title: Inpainting
-      - local: api/pipelines/stable_diffusion/depth2img
-        title: Depth-to-image
-      - local: api/pipelines/stable_diffusion/image_variation
-        title: Image variation
+      - local: api/pipelines/stable_diffusion/k_diffusion
+        title: K-Diffusion
+      - local: api/pipelines/stable_diffusion/latent_upscale
+        title: Latent upscaler
+      - local: api/pipelines/stable_diffusion/ldm3d_diffusion
+        title: LDM3D Text-to-(RGB, Depth), Text-to-(RGB-pano, Depth-pano), LDM3D Upscaler
       - local: api/pipelines/stable_diffusion/stable_diffusion_safe
         title: Safe Stable Diffusion
+      - local: api/pipelines/stable_diffusion/sdxl_turbo
+        title: SDXL Turbo
       - local: api/pipelines/stable_diffusion/stable_diffusion_2
         title: Stable Diffusion 2
       - local: api/pipelines/stable_diffusion/stable_diffusion_3
         title: Stable Diffusion 3
       - local: api/pipelines/stable_diffusion/stable_diffusion_xl
         title: Stable Diffusion XL
-      - local: api/pipelines/stable_diffusion/sdxl_turbo
-        title: SDXL Turbo
-      - local: api/pipelines/stable_diffusion/latent_upscale
-        title: Latent upscaler
       - local: api/pipelines/stable_diffusion/upscale
         title: Super-resolution
-      - local: api/pipelines/stable_diffusion/k_diffusion
-        title: K-Diffusion
-      - local: api/pipelines/stable_diffusion/ldm3d_diffusion
-        title: LDM3D Text-to-(RGB, Depth), Text-to-(RGB-pano, Depth-pano), LDM3D Upscaler
       - local: api/pipelines/stable_diffusion/adapter
         title: T2I-Adapter
-      - local: api/pipelines/stable_diffusion/gligen
-        title: GLIGEN (Grounded Language-to-Image Generation)
+      - local: api/pipelines/stable_diffusion/text2img
+        title: Text-to-image
       title: Stable Diffusion
     - local: api/pipelines/stable_unclip
       title: Stable unCLIP

docs/source/en/api/loaders/lora.md

@@ -25,7 +25,10 @@ LoRA is a fast and lightweight training method that inserts and trains a signifi
 - [`SanaLoraLoaderMixin`] provides similar functions for [Sana](https://huggingface.co/docs/diffusers/main/en/api/pipelines/sana).
 - [`HunyuanVideoLoraLoaderMixin`] provides similar functions for [HunyuanVideo](https://huggingface.co/docs/diffusers/main/en/api/pipelines/hunyuan_video).
 - [`Lumina2LoraLoaderMixin`] provides similar functions for [Lumina2](https://huggingface.co/docs/diffusers/main/en/api/pipelines/lumina2).
+- [`WanLoraLoaderMixin`] provides similar functions for [Wan](https://huggingface.co/docs/diffusers/main/en/api/pipelines/wan).
+- [`CogView4LoraLoaderMixin`] provides similar functions for [CogView4](https://huggingface.co/docs/diffusers/main/en/api/pipelines/cogview4).
 - [`AmusedLoraLoaderMixin`] is for the [`AmusedPipeline`].
+- [`HiDreamImageLoraLoaderMixin`] provides similar functions for [HiDream Image](https://huggingface.co/docs/diffusers/main/en/api/pipelines/hidream)
 - [`LoraBaseMixin`] provides a base class with several utility methods to fuse, unfuse, unload, LoRAs and more.
 
 <Tip>
@@ -77,10 +80,22 @@ To learn more about how to load LoRA weights, see the [LoRA](../../using-diffuse
 
 [[autodoc]] loaders.lora_pipeline.Lumina2LoraLoaderMixin
 
+## CogView4LoraLoaderMixin
+
+[[autodoc]] loaders.lora_pipeline.CogView4LoraLoaderMixin
+
+## WanLoraLoaderMixin
+
+[[autodoc]] loaders.lora_pipeline.WanLoraLoaderMixin
+
 ## AmusedLoraLoaderMixin
 
 [[autodoc]] loaders.lora_pipeline.AmusedLoraLoaderMixin
 
+## HiDreamImageLoraLoaderMixin
+
+[[autodoc]] loaders.lora_pipeline.HiDreamImageLoraLoaderMixin
+
 ## LoraBaseMixin
 
 [[autodoc]] loaders.lora_base.LoraBaseMixin

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

@@ -100,6 +100,8 @@ AuraFlow can be compiled with `torch.compile()` to speed up inference latency ev
 )
 ```
 
+Specifying `use_duck_shape` to be `False` instructs the compiler if it should use the same symbolic variable to represent input sizes that are the same. For more details, check out [this comment](https://github.com/huggingface/diffusers/pull/11327#discussion_r2047659790).
+
 This enables from 100% (on low resolutions) to a 30% (on 1536x1536 resolution) speed improvements.
 
 Thanks to [AstraliteHeart](https://github.com/huggingface/diffusers/pull/11297/) who helped us rewrite the [`AuraFlowTransformer2DModel`] class so that the above works for different resolutions ([PR](https://github.com/huggingface/diffusers/pull/11297/)).

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

@@ -133,6 +133,60 @@ output = pipe(
 export_to_video(output, "wan-i2v.mp4", fps=16)
 ```
 
+### First and Last Frame Interpolation
+
+```python
+import numpy as np
+import torch
+import torchvision.transforms.functional as TF
+from diffusers import AutoencoderKLWan, WanImageToVideoPipeline
+from diffusers.utils import export_to_video, load_image
+from transformers import CLIPVisionModel
+
+
+model_id = "Wan-AI/Wan2.1-FLF2V-14B-720P-diffusers"
+image_encoder = CLIPVisionModel.from_pretrained(model_id, subfolder="image_encoder", torch_dtype=torch.float32)
+vae = AutoencoderKLWan.from_pretrained(model_id, subfolder="vae", torch_dtype=torch.float32)
+pipe = WanImageToVideoPipeline.from_pretrained(
+    model_id, vae=vae, image_encoder=image_encoder, torch_dtype=torch.bfloat16
+)
+pipe.to("cuda")
+
+first_frame = load_image("https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/flf2v_input_first_frame.png")
+last_frame = load_image("https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/flf2v_input_last_frame.png")
+
+def aspect_ratio_resize(image, pipe, max_area=720 * 1280):
+    aspect_ratio = image.height / image.width
+    mod_value = pipe.vae_scale_factor_spatial * pipe.transformer.config.patch_size[1]
+    height = round(np.sqrt(max_area * aspect_ratio)) // mod_value * mod_value
+    width = round(np.sqrt(max_area / aspect_ratio)) // mod_value * mod_value
+    image = image.resize((width, height))
+    return image, height, width
+
+def center_crop_resize(image, height, width):
+    # Calculate resize ratio to match first frame dimensions
+    resize_ratio = max(width / image.width, height / image.height)
+    
+    # Resize the image
+    width = round(image.width * resize_ratio)
+    height = round(image.height * resize_ratio)
+    size = [width, height]
+    image = TF.center_crop(image, size)
+    
+    return image, height, width
+
+first_frame, height, width = aspect_ratio_resize(first_frame, pipe)
+if last_frame.size != first_frame.size:
+    last_frame, _, _ = center_crop_resize(last_frame, height, width)
+
+prompt = "CG animation style, a small blue bird takes off from the ground, flapping its wings. The bird's feathers are delicate, with a unique pattern on its chest. The background shows a blue sky with white clouds under bright sunshine. The camera follows the bird upward, capturing its flight and the vastness of the sky from a close-up, low-angle perspective."
+
+output = pipe(
+    image=first_frame, last_image=last_frame, prompt=prompt, height=height, width=width, guidance_scale=5.5
+).frames[0]
+export_to_video(output, "output.mp4", fps=16)
+```
+
 ### Video to Video Generation
 
 ```python

examples/advanced_diffusion_training/train_dreambooth_lora_flux_advanced.py

@@ -1915,17 +1915,22 @@ def compute_text_embeddings(prompt, text_encoders, tokenizers):
             free_memory()
 
     # Scheduler and math around the number of training steps.
-    overrode_max_train_steps = False
-    num_update_steps_per_epoch = math.ceil(len(train_dataloader) / args.gradient_accumulation_steps)
+    # Check the PR https://github.com/huggingface/diffusers/pull/8312 for detailed explanation.
+    num_warmup_steps_for_scheduler = args.lr_warmup_steps * accelerator.num_processes
     if args.max_train_steps is None:
-        args.max_train_steps = args.num_train_epochs * num_update_steps_per_epoch
-        overrode_max_train_steps = True
+        len_train_dataloader_after_sharding = math.ceil(len(train_dataloader) / accelerator.num_processes)
+        num_update_steps_per_epoch = math.ceil(len_train_dataloader_after_sharding / args.gradient_accumulation_steps)
+        num_training_steps_for_scheduler = (
+            args.num_train_epochs * accelerator.num_processes * num_update_steps_per_epoch
+        )
+    else:
+        num_training_steps_for_scheduler = args.max_train_steps * accelerator.num_processes
 
     lr_scheduler = get_scheduler(
         args.lr_scheduler,
         optimizer=optimizer,
-        num_warmup_steps=args.lr_warmup_steps * accelerator.num_processes,
-        num_training_steps=args.max_train_steps * accelerator.num_processes,
+        num_warmup_steps=num_warmup_steps_for_scheduler,
+        num_training_steps=num_training_steps_for_scheduler,
         num_cycles=args.lr_num_cycles,
         power=args.lr_power,
     )
@@ -1949,7 +1954,6 @@ def compute_text_embeddings(prompt, text_encoders, tokenizers):
                 lr_scheduler,
             )
         else:
-            print("I SHOULD BE HERE")
             transformer, text_encoder_one, optimizer, train_dataloader, lr_scheduler = accelerator.prepare(
                 transformer, text_encoder_one, optimizer, train_dataloader, lr_scheduler
             )
@@ -1961,8 +1965,14 @@ def compute_text_embeddings(prompt, text_encoders, tokenizers):
 
     # We need to recalculate our total training steps as the size of the training dataloader may have changed.
     num_update_steps_per_epoch = math.ceil(len(train_dataloader) / args.gradient_accumulation_steps)
-    if overrode_max_train_steps:
+    if args.max_train_steps is None:
         args.max_train_steps = args.num_train_epochs * num_update_steps_per_epoch
+        if num_training_steps_for_scheduler != args.max_train_steps:
+            logger.warning(
+                f"The length of the 'train_dataloader' after 'accelerator.prepare' ({len(train_dataloader)}) does not match "
+                f"the expected length ({len_train_dataloader_after_sharding}) when the learning rate scheduler was created. "
+                f"This inconsistency may result in the learning rate scheduler not functioning properly."
+            )
     # Afterwards we recalculate our number of training epochs
     args.num_train_epochs = math.ceil(args.max_train_steps / num_update_steps_per_epoch)
 

examples/community/pipeline_stable_diffusion_xl_controlnet_adapter_inpaint.py

@@ -33,7 +33,6 @@
 from diffusers.image_processor import PipelineImageInput, VaeImageProcessor
 from diffusers.loaders import (
     FromSingleFileMixin,
-    StableDiffusionLoraLoaderMixin,
     StableDiffusionXLLoraLoaderMixin,
     TextualInversionLoaderMixin,
 )
@@ -300,7 +299,7 @@ def rescale_noise_cfg(noise_cfg, noise_pred_text, guidance_rescale=0.0):
 
 
 class StableDiffusionXLControlNetAdapterInpaintPipeline(
-    DiffusionPipeline, StableDiffusionMixin, FromSingleFileMixin, StableDiffusionLoraLoaderMixin
+    DiffusionPipeline, StableDiffusionMixin, FromSingleFileMixin, StableDiffusionXLLoraLoaderMixin
 ):
     r"""
     Pipeline for text-to-image generation using Stable Diffusion augmented with T2I-Adapter

examples/controlnet/README_flux.md

@@ -6,7 +6,19 @@ Training script provided by LibAI, which is an institution dedicated to the prog
 > [!NOTE]
 > **Memory consumption**
 >
-> Flux can be quite expensive to run on consumer hardware devices and as a result, ControlNet training of it comes with higher memory requirements than usual. 
+> Flux can be quite expensive to run on consumer hardware devices and as a result, ControlNet training of it comes with higher memory requirements than usual.
+
+Here is a gpu memory consumption for reference, tested on a single A100 with 80G.
+
+| period | GPU |
+| - | - | 
+| load as float32 | ~70G |
+| mv transformer and vae to bf16 | ~48G |
+| pre compute txt embeddings | ~62G |
+| **offload te to cpu** | ~30G |
+| training | ~58G |
+| validation | ~71G |
+
 
 > **Gated access**
 >
@@ -98,8 +110,9 @@ accelerate launch train_controlnet_flux.py \
     --validation_image "./conditioning_image_1.png" "./conditioning_image_2.png" \
     --validation_prompt "red circle with blue background" "cyan circle with brown floral background" \
     --train_batch_size=1 \
-    --gradient_accumulation_steps=4 \
+    --gradient_accumulation_steps=16 \
     --report_to="wandb" \
+    --lr_scheduler="cosine" \
     --num_double_layers=4 \
     --num_single_layers=0 \
     --seed=42 \

examples/controlnet/train_controlnet_flux.py

@@ -148,7 +148,7 @@ def log_validation(
                     pooled_prompt_embeds=pooled_prompt_embeds,
                     control_image=validation_image,
                     num_inference_steps=28,
-                    controlnet_conditioning_scale=0.7,
+                    controlnet_conditioning_scale=1,
                     guidance_scale=3.5,
                     generator=generator,
                 ).images[0]
@@ -1085,8 +1085,6 @@ def compute_embeddings(batch, proportion_empty_prompts, flux_controlnet_pipeline
         return {"prompt_embeds": prompt_embeds, "pooled_prompt_embeds": pooled_prompt_embeds, "text_ids": text_ids}
 
     train_dataset = get_train_dataset(args, accelerator)
-    text_encoders = [text_encoder_one, text_encoder_two]
-    tokenizers = [tokenizer_one, tokenizer_two]
     compute_embeddings_fn = functools.partial(
         compute_embeddings,
         flux_controlnet_pipeline=flux_controlnet_pipeline,
@@ -1103,7 +1101,8 @@ def compute_embeddings(batch, proportion_empty_prompts, flux_controlnet_pipeline
             compute_embeddings_fn, batched=True, new_fingerprint=new_fingerprint, batch_size=50
         )
 
-    del text_encoders, tokenizers, text_encoder_one, text_encoder_two, tokenizer_one, tokenizer_two
+    text_encoder_one.to("cpu")
+    text_encoder_two.to("cpu")
     free_memory()
 
     # Then get the training dataset ready to be passed to the dataloader.