Z-Image

Quantised media packages of Tongyi-MAI/Z-Image.

MECHA Forge

MECHA — Model Engineering & Contextual Hearth Assembly

These media packages have been forged and optimized in the MECHA Forge. Each one holds the Denoiser, Text Encoder, and VAE components packaged together for easier deployment.

Description

Z-Image: undistilled foundation image model. Supports CFG (4.0), negative prompts, LoRA/ControlNet. 28-50 steps.

Available quantizations

Quant	Denoiser	Text encoder	VAE
FP4	FP4	FP4	FP8
FP8	FP8	FP8	FP8
INT4	INT4	INT4	INT8
INT8	INT8	INT8	INT8

Recommended settings

From the source model card (adjust to taste):

Parameter	Value
Steps	28-50
CFG / Guidance scale	3.0-5.0
Resolution	512-2048 (total pixels)
Notes	Supports CFG and negative prompts; good for LoRA/ControlNet

Example prompts

Caught in the moment (candid, slice of life):

masterpiece, best quality, highres, absurdres, very aesthetic, safe, 1man,
solo, candid moment, natural lighting. An open-air street cafe on a sunny summer
afternoon, a small round bistro table under a striped patio parasol, potted
plants and climbing flowers, warm golden sunlight and a gentle breezy
atmosphere. A young man with short dark hair and brown eyes. He has a lean
build. He wears a teal linen shirt with sleeves rolled up, and dark
trousers, and tan loafers. He sits at the cafe table with a glass of iced
coffee, and a relaxed smile.

Fashion photography (magazine cover style):

Fashion photography, magazine cover style: a 33-year-old woman with oval face,
brown eyes, and soft lips; wearing a floral maxi dress, straw hat, sandals;
garden backlight, elegant typography, 8K, sharp focus.

Scenery:

Scenery: a lavender field in Provence at golden hour, purple rows, stone
farmhouse, soft light, bees, photorealistic, 8K.