Verified: a year ago
SafeTensor
Type | |
Stats | 333 97 3 |
Reviews | (46) |
Published | Dec 26, 2023 |
Base Model | |
Training | Steps: 15,120 Epochs: 252 |
Usage Tips | Clip Skip: 2 Strength: 0.4 |
Trigger Words | firstSunrise,outdoors,mount fuji,mountainous horizon,sunrise,sunset,sunlight,city,ocean, roomNewYear,holding drink,english text,evening,indoors,juice,happy new year,television,holding glass,reaching towards viewer,incoming food,book,bookshelf,bottle,cabinet,cable knit,calendar \(object\),lamp,lamppost,refrigerator,remote control,table,bed,cake,(happy new year exlish text on television in background), evening,outdoors,english text,evening,happy new year,new year,looking at viewer,(shibuyaScreen,crosswalk,happy new year english text in large monitor in background), |
Hash | AutoV2 5C51EDCABA |
I wanted to welcome the New Year with my favorite character, so I created this Lora featuring the first sunrise, an indoor TV, and the large screens at Shibuya's scramble crossing as backgrounds. Originally, I intended to create a Lora displaying 'Happy New Year' with the sunrise over the peak of Mount Fuji. However, the result showed jumbled text and, instead of Mount Fuji, there were multiple sunrises. I wanted to correct this, but as I only used materials with clear text, Mount Fuji, and a single sunrise, I've come to accept this as a unique aspect of Stable diffusion.
Therefore, if anyone knows a method to create images with proper English text and a proper depiction of the first sunrise, I would appreciate your comments.