LoRA test
LoRA Strength: 0.3 Prompt: delete "full body" & add "face close-up" ペナちスタイルBのLoRAは、全て全身の画像で学習させています。ただ、思いのほか顔の描き方が上手だったのでものは試しに『寄り』のイラストにしてみました。 The dataset for Penachi Style B LoRA consists entirely of full-body images. However, since the face turned out unexpectedly well, I decided to test it with a close-up shot just to see how it looks. とっても可愛いひよこちゃん!これならスタイルLoRAとして万能に使えるかもしれません。 Hiyoko-chan turned out absolutely adorable! With results like this, it might be a highly versatile style LoRA that works for everything. また、SDXLとDiT1にも対応させています。SDXLは最近触っておらずテストする技術がないのであまり使えないかもしれません。 Also, it is compatible with both SDXL and DiT1. I haven’t messed with SDXL lately and don’t really have the setup to test it properly, so it might not perform as well there.
Parametreleri Göster
- Orijinal Prompt
- pncstylbbbbbpncpncbbbpncbbbstyl, face close-up, panorama view, lovely smile,テーマパーク、hykchnmdbkn, bangs, one side up, bob cut, black hair, black to purple gradient eyes, 150cm,
- Mod
- standard
- Boyut
- 864 x 1536
Yorumlar6