Choker Pull

Hello again! Today I decided to train my first Concept/Action/Pose model. I have not trained this type of LoRa before, so I might've not turned out the best.

I will most likely make a V2 of this model due to the low amount of images available to train.

Any feedback or tips are appreciated :)

Prompts: Choker_Pull, choker, pulling another's choker, pulling someone else, finger under choker, pulling with finger.
Negative prompts: furry (or in positive if you want), background characters, lowres, low quality, blurry, mosaic censorship, censored.
Optional prompts: 1girl, 1boy OR 2girls/boys, multiple girls/boys, collar, choker grab, hand under choker, looking at another, eye contact.

Known issues (after testing it myself):
- Despite its purpose, hands and fingers might not be anatomically accurate. Several generation attempts and/or Highres Fix might be needed to obtain the desired results.
- Struggles with chokers that have pendants.
- Struggles A LOT with "choker grabs" (aka using several fingers to hold the choker, or holding it with the full hand instead of a single finger).
- Tends to make necks longer than usual.
- The model leans towards furry due to the low variety of images used during training.