The ultimate sounding model, trained a 1000 1024x1024 images, hand picked and hand sorted, here are the thags that cn be used: https://docs.google.com/spreadsheets/d/1b-3yu0FZyU8W3hIASOMG7M1805VBd7n_g3jmDt8HN5A/edit?usp=sharing

Might handsort the training data later and train a LorA for each type of sounding, like one for dildo, one for beads, one for tentacles, etc.