我将两套服装放在一起训练，会有一些污染的状况，但可以通过提示词避免
v3.0使用anima base训练，拥有更小的模型体积
无外套：lucilla, a beautiful young woman with long purple hair fading to light blue, purple eyes, a small mole under her left eye, large silver hoop earrings, yellow bow with silver brooch, wearing a black sleeveless dress with geometric patterns and strap details, black gloves with silver jewelry
有外套：lucilla, a beautiful young woman with long purple hair fading to light blue, purple eyes, a small mole under her left eye, large silver hoop earrings, wearing a black wide-brimmed hat with starry constellation pattern and yellow ribbon, a yellow bow with silver brooch, white open jacket with blue accents, black and yellow patterned top with crisscross straps, black gloves
v1.0和v2.0使用IllustriousXL训练
v2.0在v1.0的基础上进行了优化。二者共用同一套提示词。
v2.0用了更小的dim/alpha，不仅缩减了体积还解决了小部分服装之间存在的污染问题，可以通过提示词优化