Verified: a month ago
SafeTensor
Type | |
Stats | 261 90 |
Reviews | (61) |
Published | Dec 14, 2024 |
Base Model | |
Trigger Words | chouzetsusaikawa_tenshi-chan |
Hash | AutoV2 6C2196626E |
Description
This LoRA model is trained on official art from インターネットオーバードーズ.
Requirements
To use this model, it must be paired with the pixel-based model available at Pixel Art LoRA (Noob or IL), as it was specifically fine-tuned using this base.
Usage Instructions
This LoRA employs a differential training method based on the aforementioned pixel model. Due to the original pixel-to-real pixel ratio of 1:8, it is recommended to generate images at a higher resolution, such as around 2048x2048. For best results, you can start by generating a lower-resolution image and then upscale it using image-to-image (i2i) techniques to achieve the desired level of detail and quality.