Search results

[train_sana_sprint] Custom save_hook overwrites trained transformer with frozen model during checkpointing

bugSomething isn't workingSomething isn't working

Status: Open.

#12444 In huggingface/diffusers;

· riteshrm opened on Oct 7, 2025

Support Wan2.2-Animate

contributions-welcome

help wantedExtra attention is neededExtra attention is needed

Status: Open.

#12441 In huggingface/diffusers;

· asomoza opened on Oct 6, 2025

get_1d_sincos_pos_embed_from_grid uses float64 unavailable on MPS

bugSomething isn't workingSomething isn't working

Status: Open.

#12432 In huggingface/diffusers;

· bghira opened on Oct 5, 2025

[ko] Update outdated Korean documentation in docs/source/ko 🇰🇷<ko>

Status: Open.

#12431 In huggingface/diffusers;

· braintrue opened on Oct 4, 2025

UNet-only adapters will still cause problems with pipeline.set_adapters() after torch.compile because the compiled graph expects adapters to exist in all components.

Status: Open.

#12427 In huggingface/diffusers;

· jiangtaozh opened on Oct 3, 2025

Work WorkNCCL(SeqNum=2521, OpType=ALLREDUCE, NumelIn=160672064, NumelOut=160672064, Timeout(ms)=600000) timed out in blocking wait.

bugSomething isn't workingSomething isn't working

Status: Open.

#12422 In huggingface/diffusers;

· hhy5861 opened on Oct 2, 2025

SVG 2 kernels

Status: Open.

#12415 In huggingface/diffusers;

· bhack opened on Oct 1, 2025

Validation of results with quantitative evaluation metrics - Text to Image LORA

Status: Open.

#12414 In huggingface/diffusers;

· HripsimeS opened on Oct 1, 2025

HunyuanImage-3.0 support

Status: Open.

#12412 In huggingface/diffusers;

· ryan-seungyong-lee opened on Oct 1, 2025

WanVACEPipeline does not work with low-noise stage only

bugSomething isn't workingSomething isn't working

Status: Open.

#12410 In huggingface/diffusers;

· vladmandic opened on Sep 30, 2025

Diffusers crash due to hard-dependency on <code>torch._C._distributed_c10d</code>

bugSomething isn't workingSomething isn't working

Status: Open.

#12409 In huggingface/diffusers;

· vladmandic opened on Sep 30, 2025

Generated image is cropped when using smaller output size to reduce VRAM and processing time in Flux Kontext

Status: Open.

#12406 In huggingface/diffusers;

· dangph-alala opened on Sep 29, 2025