Type | |
Stats | 2,898 3,882 |
Reviews | (475) |
Published | Jan 23, 2023 |
Base Model | |
Training | Epochs: 40 |
Trigger Words | oda eiichirou |
Hash | AutoV2 F89BC3F7F1 |
Eiichirou Oda (One Piece) style LoRA
Making models can be expensive. Do you like what I do? Consider supporting me on Patreon 🅿️ or feel free to buy me a coffee ☕
LoRA net for the style of my favourite mangaka. Oda's style has changed a lot during the years, so it was hard to capture his character design. This model mostly got his way of drawing lines and his coloring, plus certain details. It didn't really grasp character concept completely as much as my previous LoRA's (probably because the dataset was full of pics with multiple characters, as Oda mostly makes those), but it's pretty good when combined with embeddings (like the first Nico Robin pic, which is using my embedding, and unlike the second one, which is not using embeddings and looks like a mix between Nami and Robin).
Trigger word is oda eiichirou
, suggested weight around 0.5 or 0.6 depending on the base model (I used Anything V4.5 here).
For the One Piece anime style please use this other LoRA I made: https://civitai.com/models/4219/one-piece-wano-style-lora
How to use LoRA's in auto1111:
Update webui (use
git pull
like here or redownload it)Copy the file to
stable-diffusion-webui/models/lora
Select your LoRA like in this video
Make sure to change the weight (by default it's
:1
which is usually too high)