Drive-Thru Window Operator (LORA)

"Hello there, Civitai, can I take your order?"

I basically just thought of a random, unique idea of having someone as a drive-thru window operator and lo and behold, here it is. That, and when I do the prompts of someone being inside a drive-thru window, SD usually generates random locations inside the restaurant for my subjects, so I wanted to refine that. Let the hilarity ensue!

Recommended settings:

Make sure you check the hires fix (without it, the LoRA could produce very blurry results, especially with the faces)

Upscale by 2, to either 768x768 or 1024x1024 (for best quality)

Sampler: Euler a or DPM++ 2M Karras (they're the best samplers that I've used for the model since others aren't as satisfactory with it, imo)

Denoising: ~0.55-0.75

CFG scale: ~6-8

LoRA weight: ~0.6-0.9

Checkpoints: Realistic Vision, Deliberate, Dreamshaper, and Lyriel are my personal recommendations since they're the most accurate when it comes to depictions of people.

Make sure you add the keyword "inside the drive-thru window" in order for your subject to actually be inside the window. You can also add keywords of "headset", "apron", "cap hat", and "holding out (insert food here)" to make your image more creative.

Negative prompt that works well:

deformed, bad ???????, disfigured, poorly drawn face, mutation, mutated, extra limb, ugly, disgusting, poorly drawn hands, missing limb, floating limbs, disconnected limbs, malformed hands, blurry, ((mutated hands and fingers)), duplicate images, duplicate, (deformed, distorted, disfigured:1.3), poorly drawn, bad ???????, wrong ???????, extra limb, missing limb, floating limbs, (mutated hands and fingers:1.4), disconnected limbs, mutation, mutated, ugly, disgusting, blurry, amputation, EasyNegative, ng_deepnegative_v1_75t

DeepNegative and EasyNegative are very recommended.

Suggestions and feedback are welcome! I hope you enjoy!