
Lora trained in re-create the Fishinlol's style.
Illustrious V1:
I'm understanding how Illustrious works. I rlly recommend you use long descriptions or fill your prompt with commas "," until get almost 76 tokens for the positive prompt, that helps a lot to get better results. You can see some examples in the previews.
Weight: 1
Pony6 v1.5:
This one had several issues like be undercooked (in the style) and deformations. I re-trained it with other params to reduce that issues. Still doesn't need a trigger word. If you have any issue, maybe could be the model that are you using, try with Base Ponyv6, Autism or my PonyMerge pls.
Weight: .8 - 1
Pony6 v1:
For this one I had to increase the dataset with custom variations. That cause that I had to re-train over the same weights for make it works xD, could be a little burned (I mean with characters and poses) but just use it with a weight lower than 1.
Aux tag: fishlsty
Weight: .8
描述:
训练词语:
名称: fishinlol_style_pony6_v3-000038.safetensors
大小 (KB): 260171
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success