sd-scripts · Issues· 909 open

Crashed randomly in the middle of cooking

`torch.cuda.amp` in `__main__` self-test blocks and the IPEX hijack — deprecated since torch 2.3/2.4, unguarded on the main install path

[Feature Request] Add option to disable cuDNN/MIOpen backend (workaround for AMD RDNA 4 GPU crashes)

WeightsUnpickler error: Unsupported operand 56

--tag_replacement doesn't actually replace anything (variable shadowing bug)

[bug] --skip_until_initial_step 使用時、--gradient_accumulation_steps > 1 で再開位置を過剰にスキップして学習が終了する

Silent `adaln` drop while anima lora training

LECO: verify_training_args() crashes on missing sample_every_n_epochs/sample_every_n_steps

Adding Anima ControlNet support

"Missing keys in state dict" solved when training lora on some anima ft checkpoints

Can not generate full body with face lora model.

bf16/float RuntimeError - dtype incompatibility

Does Anima model support DoRA training?

Tag info saved in model metadata ignores keep_tokens_separator

Significant bottleneck with text encoder outputs and latent caching to disk with .npz vs .safetensors