This is Andrea from Star Trek The Original Series (season 1 episode 9). In the show, Andrea is a female android built sometime between 2261 and 2266, by Roger Korby.It was trained on FLUX and works great on FLUX.1 [dev], FLUX.1 [schnell], but also Acorn Is Spinning FLUX.Edit: I had to upload some different sample images as the outfit she wears in the tv show is too racy for civitai terms of service and sample images depicting them got blocked. I have a few samples from Acorn Is Spinning V1.1 Hyper 8 Step with Chin Fix (10 steps only) and to make them easy to spot, they'll be the ones with the acorn in the photo.This character was designed and produced in 1966 and the LoRA is trained from a healthy amount of screencaps from the episode. As such, it will default to a 60's cinematic look with softer image, but you can easily prompt it for a clear photo. That said, FLUX will want to generate images with backgrounds that look like they were taken in the 1960's. I think this is pretty cool as I prompted for a Starbucks coffee shop background and it gave me a 1960's Starbucks. If you reference things from different time periods, I'm sure FLUX will figure that out too. Another example is where there is a sports car at a dealership. If I didn't specify modern sports car, then I got 1960's sports cars. However, just throwing in modern day sports car got me a new ferrari in the image. To the hardcore Trekkies out there - I apologize for not taking careful consideration into getting her outfit just right. You will see what I mean . FLUX will change the pattern and style a little bit, but it gets the overall look and knows to get the top part mostly right. I blame the screencaps, but if it bothers anyone that much, I could probably clean up the dataset and mitigate the deviations with careful captions. There's a little bit of Kirk in there, but I found that if I tried to tag and prompt for him, there just wasn't enough in the dataset to avoid generating closeups with her eyes crossed.Let me know what you think or if you can think of any other characters worthy of this sort of thing. I honestly didn't think it would turn out as well as it did given the limited dataset (the whole episode pretty much takes place in 3 or 4 rooms/sets).

Files with SHA256: 374667c9b3205c8e971ec753cd621c392078f826e72a155ebca33e9027b8567e