ds7485

Curious if anyone else has noticed that the text-to-image function has gotten worse recently. I used to find it pretty great at generating images in a particular style - landscapes that looked like actual paintings, etc., provided I phrased my prompt specifically - but now everything seems to trend towards the "digital painting" style, and photos again look like cut-and-paste collages. I've only been using the default (https://perchance.org/ai-text-to-image-generator) so if this issue is confined to this generator, great. Any insight would be helpful!

Hi, everyone Does anyone happen to know how the default character generator creates names, and if that's editable? I know it's a pretty trivial thing - I can come up with my own easily enough - but I get a little tired of the same names being used over and over (and over, and over.) Is this something I can tweak on my end, or is it embedded pretty deep in the text-generation plugin(s)?