Choco Commission

It's finally done! I am sorry about the lack of updates these past few days; I was just so tired of delaying this set so I resolved to not work on anything else until its done. If I said "Choco should be up tomorrow" and it wasn't then I would've gone crazy. It almost felt really close to done so every day I was expecting to finish it, but I ran into some more snags at the finish line. Anyways, I hope you like it! I hope I captured Choco's personality and slammin' bod.

I really need to post more to make up for this dead period of time and to get everything I want done this month! Next up is more commissions and I'll mix Kobo/Marine in there. I also baked a new lora for Marine's sololive outfit and it's a lot better than the old one I did, so I can use that for a set too.

I owe an explanation as to why it took this long. To start with the commissioner let me cook of which I am grateful of; all the parts of the set which made it take so long were my own ideas and entirely my fault. I made some decisions early on of which I would not realize the consequences of until later.

• The background obviously. It seemed like a classic setting which would be ideal for the set, but the desks created a lot of scuff and headache.

• I felt the lighting was pretty dull so I added a sunset theme. The sunset certainly spices up the environment; but the AI wants the sun to be visible at all times, so keeping the windows roughly consistent turned to be tough and the sun's position made any outputs awkward. In the future I probably won't do sunsets indoors lol.

• I had an idea to do the set out of order. Do all the tough parts first and then connect them, so for example number 8 was the second image I did. This was cool until I realized I needed to get Choco off the desk and on to the floor. Figuring that transition was tougher since I'd already committed to some action in both places and my various ideas didn't work. I eventually just prompted through it but that was the roadblock I referred to a few times. The sunk cost fallacy is real!
  
  All of that combines to just having a low hit-rate overall in my batches so it was a lot of prompting to find the right image. My takeaway from this is all of the improvements to speed do not matter if the hit-rate is low overall, so prompt selection is still very important. Regardless I had a lot of fun actually putting together the images that I did hit; I think they have a lot of character to them! Let me know what you think!