Verified: 2 months ago
SafeTensor
Type | |
Stats | 280 3,200 |
Reviews | (70) |
Published | Dec 20, 2024 |
Base Model | |
Training | Steps: 12,810 Epochs: 30 |
Usage Tips | Clip Skip: 2 Strength: 1 |
Trigger Words | by ajimita by ame929 by ashima \(roro046\) by ask \(askzy\) by dino \(dinoartforame\) by diyokama by dsmile by fuzichoco by kinta \(distortion\) by konya karasue +16 more |
Hash | AutoV2 B095CAFC81 |
Artist Collection - Study
Collection of 25 curated and manually rated artist datasets:
ajimita
ame929
ashima \(roro046\)
ask \(askzy\)
dino \(dinoartforame\)
diyokama
dsmile
fuzichoco
kinta \(distortion\)
konya karasue
lack
lam
lifeline \(a384079959\)
mika pikazo
miv4t
mizumi zumi
nid417
quasarcake
ruoganzhao
say hana
timbougami
tokkyu
wanke
xixun
yoneyama mai
z3zz4
These artists tags are already trained in Illustrious/noobai, (see x/y/z plot references) this is an attempt to further train on specifc works to appreciate and study their styles.
Recommended prompt structure:
Positive prompt:
by {artist},
{{tags}}
masterpiece, best quality, very awa, absurdres
Negative prompt:
(worst quality, low quality:1.1), bad anatomy, bad hands, watermark, ugly, distorted, signature
I trained the artists datasets together to experiment with artist tag combinations and prompting techniques like alternating words for swapping on each step:
by [say hana|wanke|diyokama|yoneyama mai]