| license: apache-2.0 | |
| base_model: Qwen/Qwen-Image-2512 | |
| pipeline_tag: text-to-image | |
| library_name: mlx-gen | |
| tags: | |
| - mlx | |
| - mlx-gen | |
| - mflux | |
| - apple-silicon | |
| - 8-bit | |
| - qwen | |
| - qwen-image | |
| # qwen-image-2512-8bit | |
| This repository contains MLX-Gen saved weights for `Qwen/Qwen-Image-2512`. The checkpoint is designed for local Apple Silicon inference with [`mlx-gen`](https://github.com/lpalbou/mlx-gen). | |
| It uses the mflux/MLX saved-weight layout and MLX quantization tensors. It is not a Diffusers or Transformers `from_pretrained()` checkpoint. | |
| ## Source Model | |
| Original model: [`Qwen/Qwen-Image-2512`](https://huggingface.co/Qwen/Qwen-Image-2512). | |
| ## License and Access | |
| This quantized derivative follows the Apache 2.0 license of the source model. | |
| ## Quantization | |
| This is an MLX q8 checkpoint. Quantizable modules are saved at 8-bit where the model layout supports MLX quantization; VAE weights and non-quantizable layers remain BF16. The Qwen-specific mixed q4/q8 policy only applies when preparing Qwen models with `--quantize 4`; see the [MLX-Gen quantization docs](https://github.com/lpalbou/mlx-gen/blob/main/docs/quantization.md). | |
| ## Compatibility | |
| Requires `mlx-gen >= 0.18.2`. | |
| Generated with `mlx-gen 0.18.2`. | |
| Use the `mlxgen` command and Python import path for new MLX-Gen projects. | |
| ## Usage | |
| ```bash | |
| python -m pip install -U mlx-gen | |
| mlxgen download --model AbstractFramework/qwen-image-2512-8bit | |
| mlxgen generate \ | |
| --model AbstractFramework/qwen-image-2512-8bit \ | |
| --prompt "Your prompt here" \ | |
| --steps 20 \ | |
| --seed 42 \ | |
| --output image.png | |
| ``` | |
| ## Attribution | |
| MLX-Gen is based on [mflux](https://github.com/filipstrand/mflux) by Filip Strand and the original mflux contributors. This model card is generated by MLX-Gen so derived checkpoints keep that attribution visible. | |
| Quantized and contributed by [@lpalbou](https://huggingface.co/lpalbou). | |