Тип
TextualInversion
Время публикации
2025-01-15
Базовая модель
SD 1.5
Слово-триггер
Any4T4yl0rJ0y_v2_HM-142
Копировать
Версия введения.
Another retrain instead of a second step. New captions, lower learning rate, higher gradient accumulation steps. Slightly further in the steps to find a good version, but I think it's a bit more accurate and flexible.

