Type | |
Stats | 66 1,808 |
Reviews | (23) |
Published | Feb 1, 2025 |
Base Model | |
Usage Tips | Clip Skip: 2 Strength: 1 |
Trigger Words | waffletc furry male, dog boy, grey eyes, brown hair, animal ears, animal nose, whiskers, tail, body fur police uniform, police hat, boots, gloves, short sleeves, pants, red necktie, badge |
Hash | AutoV2 D5F202085F |
Waffle Ryebread is the main protagonist of the PS1 game "Tail Concerto." The 1st game from the "Little Tail Bronx" series from CyberConnect2. He also appears in the second Little Tail Bronx game "Solatorobo: Red the Hunter"
Illustrious:
Changed up the dataset to work better for SDXL.
Weight: 1
Triggerword: waffletc
Useful tags: furry male, dog boy, grey eyes, brown hair, animal ears, animal nose, whiskers, tail, body fur
Outfit: police uniform, police hat, boots, gloves, short sleeves, pants, red necktie, badge
His hat works better from behind this time but often doesn't include the emblem thing, which is for the best because when it is there, it almost always looks terrible. His arm patch thing is also a bit crusty, but can sometimes turn out okay when close-up.
His hair was captioned with "brown hair" and "short hair" but for some reason, it sometimes gives him more hair or incorrect hair to what he's supposed to have. With some testing, I found that using just "brown hair" with lower weights works often. example: (brown hair:0.4). Artist tags seem to effect this greatly.
SD 1.5:
Trained on Indigo V105
Weight: 1
Triggerword: waffletc
Useful tags: 1boy, furry male, dog boy, grey eyes, brown hair, short hair, animal ears, animal nose, whiskers, tail, body fur
Outfit: police uniform, police hat, boots, gloves, short sleeves, pants, red necktie, badge
I tried to make a Pony version of this over half a year ago and it was awful. Bigger dataset, smaller dataset, changed captioning... no matter what, Waffle always looked like he was put through a piss filter and looked terrible. Trained more, trained less, lower weights.... Nothing worked. Every epoch in every attempt SUCKED. After several attempts, I gave up since I had no idea what was wrong. Never had anything else have this issue in Pony. Didn't do anything different from what I normally did with my other LoRAs. The dataset worked well for Illustrious with no issues, so I'm just gonna say that Pony Diffusion hates me since it's the only reason I could come up with... Though the real reason is probably that I kept doing something wrong at the time ¯\_(ツ)_/¯