Updated: Nov 12, 2024
styleVerified: 7 months ago
SafeTensor
Type | |
Stats | 1,867 68,586 |
Reviews | (363) |
Published | Apr 7, 2024 |
Base Model | |
Training | Steps: 4,200 Epochs: 10 |
Usage Tips | Clip Skip: 2 Strength: 0.8 |
Hash | AutoV2 A1B3DFB3DD |
LoRa based SunsetRiders7 art-style.
Advices:
Use "score_9, score_8_up, score_7_up, score_6_up, score_5_up" at start of the prompt.
Use AutismMix Models for better results (However, v2.0 work good with Pony and PDXL)
Recomended Strenght: 0.8-1.0, but feel free to experiment!
Inpaint faces with 0.25-0.4 denoising strenght for best results!
Negative prompt: "simple background" - because original set have a lot of flat colored backgrounds.
v1.0 - based only on faces, so I guess its better for them, inpaint faces with this can make better results.
v2.0 -based with large set of full images with poses, so its work better with full scenes, exceptionally a naughty ones.