Stable Diffusion · LoRA

Snapchat Selfie Redmond Z-Image LoRA

LoRA for Tongyi-MAI/Z-Image-Turbo trained on Snapchat selfie images to generate consistent, phone-camera-style selfies.

About this model

This LoRA adapts Tongyi-MAI/Z-Image-Turbo toward the Snapchat selfie aesthetic: casual, phone-camera-style portraits with a candid, social-media feel. Per the model card, it was trained on Snapchat Selfie style images, with GPU time provided by Redmond.AI, and the author describes the outputs as high-quality with excellent consistency across generations.

The sample images confirm the style: subjects photographed at arm's length in everyday indoor settings — a bedroom, a study with bookshelves, a gym mirror — with the phone in hand and the loose framing of a real selfie rather than a posed studio portrait. Trigger phrases are "Snapchat Selfie" and "Snapchat pic." The weights ship in safetensors format, and the author recommends running the LoRA in ComfyUI on top of the Z-Image Turbo base model.

The model is released under the Apache License 2.0. The card does not document recommended CFG scale or step counts, so follow your usual Z-Image Turbo settings and tune the LoRA weight for your prompt.