It can draw vintage poster style images of women, men, landscapes and whatnot. i think the model can create very interesting content.
Please see the preview images, note that:
1 It's expected to see gibberish text here and there - it's a fundamental part of the training images.
2 The colors will most likely be very bright.
3 A very small portion of the training images was NSFW - so even if the risk is low I flagged this.