diffusers

diffusers copied to clipboard

[LoRA training] add aspect ratio bucketing

3

support multi-aspect training

linoytsaban

Add SkyReels V2: Infinite-Length Film Generative Model

19

Thanks for the opportunity to fix #11374! Original Work Original repo: https://github.com/SkyworkAI/SkyReels-V2 Paper: https://huggingface.co/papers/2504.13074 > **SkyReels V2's main contributions are summarized as follow: • Comprehensive video captioner that understand the...

tolgacangoz

[wip][lora]feat: use exclude modules to loraconfig.

1

# What does this PR do? If we try to do: ```py from diffusers import AutoModel, DiffusionPipeline import torch model_id = "Wan-AI/Wan2.1-VACE-14B-diffusers" vae = AutoModel.from_pretrained(model_id, subfolder="vae", torch_dtype=torch.float32) pipe = DiffusionPipeline.from_pretrained(model_id,...

sayakpaul

Support dynamically loading/unloading loras with group offloading

3

Fixes #11791. reproducer ```python import torch from diffusers import HunyuanVideoPipeline, HunyuanVideoTransformer3DModel from diffusers.utils import export_to_video from diffusers.hooks import apply_group_offloading from diffusers.utils.logging import set_verbosity_debug set_verbosity_debug() model_id = "hunyuanvideo-community/HunyuanVideo" transformer = HunyuanVideoTransformer3DModel.from_pretrained(...

a-r-r-o-w

[wip] Higgs quantization

1

# What does this PR do? Testing code: ```py from diffusers import DiffusionPipeline, AutoModel from diffusers.quantizers.quantization_config import HiggsConfig import torch ckpt_id = "black-forest-labs/FLUX.1-dev" transformer = AutoModel.from_pretrained( ckpt_id, subfolder="transformer", quantization_config=HiggsConfig(modules_to_not_convert=["x_embedder"]), torch_dtype=torch.bfloat16,...

sayakpaul

**Is your feature request related to a problem? Please describe.** Not a problem but a good feature to be supported in Wan Vace pipeline as model supports it. **Describe the...

nitinmukesh

device_map can not use xla device such as tpu

1

### Describe the bug ```import torch, pathlib, builtins, transformers, diffusers, huggingface_hub model_id = 'Wan-AI/Wan2.1-T2V-14B-Diffusers' model_dir = huggingface_hub.snapshot_download(model_id) pipe = diffusers.WanPipeline.from_pretrained(model_dir, vae=diffusers.AutoencoderKLWan.from_pretrained(model_dir, subfolder='vae', torch_dtype=torch.float32), torch_dtype=torch.bfloat16, device_map='balanced') pipe.scheduler = diffusers.UniPCMultistepScheduler(prediction_type='flow_prediction', use_flow_sigmas=True, num_train_timesteps=1000,...

chaowenguo0101

# What does this PR do? In accordance with `transformers`'s [recent move](https://x.com/LysandreJik/status/1933201171130593530) and also something we keep seeing (i.e., none-to-minimal usage of the said frameworks at least from the library),...

sayakpaul

fix deprecation in lora after 0.34.0 release

1

# What does this PR do? Fix https://github.com/huggingface/diffusers/actions/runs/15854526414/job/44696447831?pr=11800#step:6:2534 So that https://github.com/huggingface/diffusers/pull/11800 can be merged.

sayakpaul

[chore] post release v0.34.0

2

sayakpaul