LyCORIS

LyCORIS copied to clipboard

LoHa/LoKr with conv. Error with tensor size mismatch

1

Was testing conv models and did one with LoHa and errored about a size mismatch. Using Kohya. Commit: daa559fbd1e095d05619463d72f7f5ae0bd2a493 on dev branch Relevant parts of the config: ```toml pretrained_model_name_or_path="runwayml/stable-diffusion-v1-5" mixed_precision="fp16"...

rockerBOO

Lycoris extreme slow training after update to latest version

11

After updating to the latest release Lycoris/LoCon (the only one i tested) training in kohya_ss is super slow. With my 4090 with the same setting loaded, in the previous version...

killerciao

LyCORIS Dylora AttributeError apply_max_norm_regularization

3

I am having trouble running the LyCORIS Dylora adaption algorithm on an `SD v1.5` checkpoint in `Kohya_ss GUI v23.0.15`. I am using this on a Windows 11 with RTX 3070....

SmirkingKitsune

DinoV2 similarity score calculation

9

Dear authors, Thank you for the amazing works, could you give the implementation of Subject Fidelity usign DINOv2 and Prompt fidelity? I'd like to replicate the paper results. (https://arxiv.org/pdf/2309.14859) Once...

Sundragon1993

diag_oft/Boft [rank0]: RuntimeError: dictionary changed size during iteration

1

lycoris-lora==3.0.1.dev1 Got a runtime error when using either Diag_OFT or BOFT. `[rank0]: RuntimeError: dictionary changed size during iteration` From ``` 2024-07-27 14:41:44|[LyCORIS]-WARNING: constrain is deprecated. Please use constraint instead. [rank0]:...

GenocidalStudios

Error Initializing t2 using `weight_gen` functinal API.

2

We get an error "AttributeError: 'NoneType' object has no attribute 'shape'" when Initializing t2 parameter for Linear layer. This precisely happens because default value of t2 is None and for...

yaswanth19

DoRA implementation differs from PEFT DoRA

9

@sayakpaul and I investigated an issue with loading a LyCORIS LoRA checkpoint which uses DoRA in diffusers. For some reason, we couldn't get the shapes of the DoRA scale vector...

BenjaminBossan

GLoRA inference fails with Flux due to weights being in bfloat16 precision

1

My simple inference script is failing when calling wrapper.merge_to() with Flux Dev as the base model. ``` 2024-09-21 19:27:53|[LyCORIS]-INFO: Loading Modules from state dict... 2024-09-21 19:27:54|[LyCORIS]-INFO: 504 Modules Loaded Traceback...

mhirki

Thank you for this useful library! I noticed that the library configures logging internally by setting logger level/handlers inside the module. While this may help with development or debugging, it...

Jannchie

publish a wheel

3

I notice that this package publishes only source distributions, so that every user has to build the wheel themselves. Better for package owners to build and publish wheels once and...

dimbleby