Onyx Z Image — 3-bit MLX
A 3-bit MLX quantization of the official Tongyi-MAI/Z-Image, built for
on-device inference in the Onyx app with a documented, reproducible pipeline.
- Source: official
Tongyi-MAI/Z-Image(diffusers), quantized in-house. - Quantization: 3-bit, group 64, MLX. Only transformer + text-encoder Linear leaves are quantized;
norms/embeddings/convs and the whole VAE stay fp16. Components:
transformer/,text_encoder/(Qwen3),vae/(16-ch FLUX-family),tokenizer/,scheduler/. - Inference: ~22-28 steps + real guidance (undistilled). Built as MIXED 3/4-bit: quant-sensitive layers (embedders, final layer, adaLN modulation, attention) stay 4-bit; the FFN bulk is 3-bit. Uniform 3-bit mottles the undistilled DiT; this recovers a clean image at near-3-bit size.
- Verified: reloaded through the Onyx engine and render-checked before publish.
Quantized derivative for efficient on-device use; all credit for the model to Tongyi-MAI (Alibaba).
- Downloads last month
- 39
Hardware compatibility
Log In to add your hardware
Quantized
Model tree for wabibito/Onyx-Z-Image-3bit
Base model
Tongyi-MAI/Z-Image