---
license: mit
base_model: inclusionAI/Ming-Image-0.1-Design
library_name: mflux
pipeline_tag: text-to-image
tags:
  - mlx
  - mflux
  - apple-silicon
  - text-to-image
  - graphic-design
  - rgba
---

# Ming-Image-0.1-Design — MLX (8-bit DiT, 8-bit text encoder)

An MLX conversion of [inclusionAI/Ming-Image-0.1-Design](https://huggingface.co/inclusionAI/Ming-Image-0.1-Design)
for [mflux](https://github.com/filipstrand/mflux) — 25.2 GB in total.
Ming-Image is a design-focused text-to-image model (posters, cards, UI, typography) that
outputs **RGBA** images.

| Component | Precision | Size |
|---|---|---|
| Ling-mini-2.0 MoE text encoder (`mllm/`) | 8-bit (routers, norms, query tokens full precision) | 17.0 GB |
| Qwen2 connector (`connector/`) | 8-bit | 1.4 GB |
| Projection heads (`mlp/`) | bf16 | 0.06 GB |
| DiT (`transformer/`) | 8-bit | 6.5 GB |
| RGBA VAE (`vae/`) | bf16 | 0.25 GB |

The checkpoint's Qwen2.5 ViT, `lm_head` and audio router are not used for text-to-image and
are not included.

## Usage

Requires mflux with Ming-Image support (the `ming-image` model family; until it is merged upstream,
install mflux from the branch that adds it).

```sh
mflux-generate-ming \
  --model path/to/this/repo \
  --prompt "A modern tech conference poster titled 'MLX SUMMIT 2026' ..." \
  --width 1024 --height 1024 --seed 42
```

Defaults follow the official pipeline: 12 steps, guidance 1.0, flow-match Euler with the
checkpoint's static shift of 6. PNGs keep the alpha channel (`--flatten-alpha` for RGB).

## Performance

The DiT stage is identical to the 5-bit-text-encoder variant: on a base M4 Mac mini (24 GB) 512² takes ~1 min
(9.8 GB peak) and 1024² 4–4.5 min (14.5 GB peak). The text encoder is freed after encoding the prompt.

## Fidelity

The MLX implementation was checked stage by stage against tensors dumped from the official
PyTorch pipeline: DiT, VAE and connector at cosine similarity ≥ 0.9999, and the text encoder
within the official model's own eager-vs-flash-attention spread (this includes reproducing
the official pipeline's autocast-bf16 MoE routing exactly). From the same initial noise it reproduces the official images' layout, typography and text. The text-encoder stage needs roughly 18.5 GB (estimated from the weight sizes), so use a Mac with 32 GB or more.

The DiT is deliberately kept at 8 bits: at 4 bits large headline lettering degrades.

## License

MIT, as the original model. Credit to inclusionAI for
[Ming-Image](https://github.com/inclusionAI/Ming-Image).