Some LoRAs
Images are generated with the keyword, and basic positive/negative prompts as recommended by the model's creator. The medium strength LoRA is used and weight is set to 1, stronger and weaker LoRA are also provided.
Hero images are cherry picked with a few additional prompts. Sample images are randomly generated without cherry picking. This is to show the inheritant biases of the LoRA.
Various models are used such as PastelMix, ColorBox, AbyssOrangeMix, Corneos7thHeavenMix, Anything, EveryoneMix, Counterfeit.
Embeddings can be found here https://huggingface.co/soknife/embeddings
irys-classic-subject
Keyword: irys \(hololive\)
irys-classic-subject.safetensors
Random sample
holo-en-outfit
Keyword: hololive english, idol clothes
Random sample
zeta-subject
Keyword: vestia zeta
Keyword: kimono, japanese clothes, vestia zeta
Random sample
suisei-regular-subject
Keyword: hoshimachi suisei
suisei-regular-subject.safetensors
Random sample
ohisashiburi-style
Keyword: ohisashiburi
ohisashiburi-style.safetensors
Random sample
fauna-regular-subject
Keyword: ceres fauna
fauna-regular-subject.safetensors
Random sample
irys-regular-subject
Keyword: irys \(hololive\)
irys-regular-subject.safetensors
Random sample
Inference Providers
NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API:
The model has no library tag.