SDLearner

I've been playing around with flux for a while now. Unlike SD, it likes natural prompts.eg: If I wanted to render a bar decorated for Halloween theme, I can write down a prompt in natural language and it will get it rendered for mePrompt#1 - A spooky Halloween-themed bar, captured by a Fujifilm X-T4 DSLR, exudes an eerie ambiance. Inspired by the macabre illustrations of Edward Gorey, the dimly lit space is adorned with cobwebs, carved pumpkins, and flickering candles that cast haunting shadows across the room. The bartender, dressed in a ghoulish costume, mixes cocktails with a sinister flair, as patrons sip their drinks beneath the watchful gaze of a towering, animatronic vampire. The Nikon D850's shallow depth of field draws the viewer's attention to the bar's intricate details, from the ornate, gothic-inspired decor to the bartender's mesmerizing movements, creating an immersive, otherworldly experience.Another point to note is that it does an amazing job in very less steps. I tend to use 10-12 steps for my renders