impresstation-zimage

zimage LoRA "impresstation-zimage" — trigger "stationthis".

Trigger word: stationthis

Sample Outputs

stationthis, stationthis, low poly playstation screenshot style, A determined young woman…	stationthis, stationthis, low poly playstation screenshot style, Under a misty, twilight…
:---:	:---:
stationthis, stationthis, low poly playstation screenshot style, A stoic, sunglasses-wear…	stationthis, stationthis, low poly playstation screenshot style, A stern, determined woma…
:---:	:---:
:---:	:---:

Usage (Diffusers)

import torch
from diffusers import DiffusionPipeline

pipe = DiffusionPipeline.from_pretrained(
    "Tongyi-MAI/Z-Image-Turbo", torch_dtype=torch.bfloat16
).to("cuda")
pipe.load_lora_weights("ms2stationthis/impresstation-zimage")
image = pipe("stationthis, a character portrait", guidance_scale=4.0, num_inference_steps=25).images[0]
image.save("out.png")

Recommended Settings

LoRA strength	Guidance	Steps	Resolution
0.8–1.0	4.0	25	1024×1024

Training Details

• Base: Tongyi-MAI/Z-Image
• Steps: 3000 · Network: LoRA rank 32 / alpha 32
• Optimizer: adamw8bit, lr 1e-4 · Scheduler: flowmatch
• Resolution: 512, 768, 1024 (multi-res bucketed) · Precision: bf16 train / fp16 save

Reproduction

This repo includes the full dataset/ (33 image-caption pairs) and the exact config.yaml so the LoRA can be retrained as-is.

About

Trained on StationThis — an AI creative platform powered by $MS2. Train your own LoRAs via @stationthisbot on Telegram.

_{Published via noema.}