Flux text2img 使用 Lora

SeaArt 官方模型加载器已经支持FLUX标签和筛选，所以之前对于怎么找flux checkpoint和lora的内容就不再赘述。
目前flux是支持多lora的，但为避免应用界面过于复杂，我仅开放了一个lora位，另外默认加载了FaeTastic Details和mjV6，权重均为1。
单图的算力消耗约30PT，2-4图则会降低单图消耗，最低能到19PT每张图。建议增加出图的batch size获得最高的性价比。
SeaArt's official model loader already supports FLUX tags and filtering, so I won't go into detail about how to find flux checkpoints and lora before.
現在, flux supports multiple lora, but in order to avoid the application interface being too complicated, 我仅开放了一个lora位, and the FaeTastic Details and mjV6 are loaded by default, 权重为1.
The computing power consumption of a single image is about 30PT, and the consumption of a single image will be reduced by 2-4 images, and the minimum can be 19PT per image. It is recommended to increase the batch size of the plot to obtain the highest cost performance.