I’m using a Flux model (chroma). Quick question: If a Lora has a keyword, does it matter where in the positive prompt you put it?
I seem to remember reading somewhere that for SDXL and SD1.5 models, keyword placement did matter.
Thanks1
1 Like
Whether to place trigger words before or after, or within the context, is often important in FLUX. Generally, if the trigger word is either at the beginning or the end, it will have an effect…
However, LoRA is similar to modifying the model itself, so in extreme cases, only the creator knows the ideal usage. Rather than focusing on FLUX, it is important to read the explanations for each LoRA carefully.
There are also LoRA that do not require trigger words (e.g., those that fix the art style).
Was this the same post as yesterday?. I remember it was asking for a way to trigger semantic vectors.
1 Like