Lora favors indoor images do to training source, however is flexible enough to do outdoor, and ??? on back which was not part of the source training.

No trigger words are needed at 0.75-1.0 lora values.