Skip to content

Actions: huggingface/diffusers

Actions

Issue Labeler

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
issue_labeler.yml
432 workflow runs
432 workflow runs

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
Per-token FlowMatch Euler steps use the opposite velocity sign Issue Labeler #436: Issue #15185 opened by davrens101
21s
21s
flash_varlen raises when requesting log-sum-exp Issue Labeler #435: Issue #15182 opened by 0z5a
15s
15s
flash_4_hub does not forward return_lse to its kernel Issue Labeler #434: Issue #15180 opened by 0z5a
9m 50s
9m 50s
Gradient clipping is a no-op in train_text_to_image_lora.py and train_lcm_distill_lora_sdxl.py Issue Labeler #433: Issue #15178 opened by gorkemgul
18s
18s
img2img: outputs paired with the wrong init image Issue Labeler #432: Issue #15176 opened by Arnav66692
26s
26s
Cosmos 3 docs: FP8 snippet fails at the first W8A8 step without importing ModelOpt's FP8 GEMM backend Issue Labeler #431: Issue #15170 opened by piotrm-nvidia
26s
26s
Cosmos 3 FP8: wrapping pipe.transformer (e.g. torch.compile) silently disables the checkpoint's W8A16 step policy Issue Labeler #430: Issue #15169 opened by piotrm-nvidia
20s
20s
UniPC and DPM-Solver with use_flow_sigmas condition the model on integer timesteps while integrating with float sigmas Issue Labeler #429: Issue #15167 opened by LiRunGuo
18s
18s
[training scripts] --logit_mean biases timesteps in the opposite direction to the SD3 paper (positive m oversamples low noise) Issue Labeler #428: Issue #15166 opened by Andreas-Hum
16s
16s
Flux and LTX pipelines do not forward the generator to stochastic scheduler steps Issue Labeler #427: Issue #15163 opened by JohnLyu2
18s
18s
No LoRA fine-tuning example for RFD4-Proteina Issue Labeler #426: Issue #15162 opened by kierandidi
15s
15s
DreamLite: disabling attention slicing replaces the GQA processor and breaks inference Issue Labeler #425: Issue #15161 opened by gorkadabo1
15s
15s
Kandinsky 6: add context-parallel (_cp_plan) support to Kandinsky6Transformer3DModel Issue Labeler #424: Issue #15156 opened by vladislav-vasilenko
19s
19s
Clarify set_adapters() adapter_weights documentation and type hints Issue Labeler #423: Issue #15146 opened by gorkadabo1
18s
18s
flux_lora_quantization example: compute_embeddings.py crashes on current transformers Issue Labeler #422: Issue #15145 opened by kahlun
19s
19s
DDPM schedulers add fresh noise on terminal transitions when the final timestep is nonzero Issue Labeler #421: Issue #15139 opened by surgan12
23s
23s
[Feature Request] Add DiCache for sample-adaptive caching in FLUX transformers Issue Labeler #420: Issue #15134 opened by BlazeSR10hawk
20s
20s
[Feature] Reuse Ulysses shard lengths across MiniMax-H3 blocks Issue Labeler #419: Issue #15132 opened by LeslieWylie
20s
20s
Flux2KleinInpaintPipeline: padding_mask_crop pastes a center crop of the result instead of resizing it Issue Labeler #418: Issue #15131 opened by hkang4628
17s
17s
set_adapters applies component weights only to the first adapter when broadcasting a dictionary Issue Labeler #417: Issue #15127 opened by gorkadabo1
22s
22s
@dataclass on quant configs breaks __eq__/__repr__/hashing, and to_json_string() raises for most configs Issue Labeler #416: Issue #15126 opened by OnePunchMonk
16s
16s
Nunchaku AWQ config accepts any group_size, but the kernel only supports 64 Issue Labeler #415: Issue #15125 opened by OnePunchMonk
17s
17s
Quantizer configs are mutated in place, leaking modules_to_not_convert across loads Issue Labeler #414: Issue #15124 opened by OnePunchMonk
15s
15s
from_single_file forces pre_quantized=True, breaking bnb/TorchAO/Quanto single-file loading Issue Labeler #413: Issue #15123 opened by OnePunchMonk
16s
16s
NVIDIAModelOptConfig.to_dict() crashes from_pretrained/save_pretrained when forward_loop is set Issue Labeler #412: Issue #15122 opened by OnePunchMonk
17s
17s