Works best on 4mp resolution. Captions are not best yet, but working on a v1.5. general style, look and feel can be done with this model, but tends a bit to give men data skin. The dataset was created from 1-3 screencaps per episode of the full series, so around 400 images
Show more
The purpose of this model is to recreate the overall cinematic style of Star Trek The next generation.
K2: Trained on nearly 400 screencaps from the series.
F1,XL,Pony: Trained on 40 screencaps from the series.