Disclaimer: Recommended for offline generation.
For the most part this is not a LoRA you can just enable and brainlessly generate relying on a trigger word to sculpt the output. Due to the nature of this LoRA fighting against the base model you need to be extra careful with your positive and negative prompt. Reference the showcase images for examples of how prompting is done with the LoRA. Due to chaotic outputs, I'm putting this in here as a warning so that no one uses yellow buzz for on-site generation and feels like they wasted it.
This LoRA works for both SFW and NSFW outputs.
Draft version of an Anima remake of what was my most popular concept LoRA in my old account. If you happen to be familiar with that old LoRA, don't expect the same quality in the draft version as what was available in that Illustrious version.
This also combines the "Spyview" concept LoRA within it, which I had not managed to do before the old account was disabled. (A voyeuristic camera perspective).
Due to the merging of multiple concepts It will take a lot of fine-tuning to get the quality up. Right now it works, but not quite to the extent I am aiming for.
Description
Heavily improved the voyeur perspective.
Not recommended to use both voyeurview and fixed perspective show. You can mix them if you want to test, but outputs may be chaotic. It's fine for offline generation when it doesn't cost anything to create an image, but you are likely going to waste buzz if you try online.
When using voyeurview, you want to also include a pov reference. The following are examples:
pov doorway (most well trained)
pov curtains
pov window, from outside
Make sure to be careful not to include contrasting/conflicting tags, and also include some natural language to help sculpt the output how you want it.
Due to recent changes in CivitAI I must monetize this, otherwise I cannot refine it further or release new LoRA in the future.