intel-ai/HunyuanImage-3.0-Instruct-Distil-MXFP8-AutoRound

Model Details

This model is an MXFP8 (W8A8, group_size 32) model of tencent/HunyuanImage-3.0-Instruct-Distil generated by intel/auto-round using RTN. Please follow the license of the original model.

• Quantized (MXFP8): linear layers
• Kept in BF16: router, vision tower, embeddings, norms

How to Use

Environment

pip install vllm==0.29.0
uv pip install git+https://github.com/vllm-project/vllm-omni.git@1c7476ec19899f1e61838e23e9fad3505403b9b9
uv pip install flashinfer-python==0.6.18

vLLM Usage

vllm serve intel-ai/HunyuanImage-3.0-Instruct-Distil-MXFP8-AutoRound \
    --omni

Evaluation

Benchmark	BF16	MXFP8	Recovery
Single Object	96.25	98.44	102.28%
Two Object	87.12	84.60	97.11%
Counting	65.62	69.69	106.20%
Colors	88.03	85.64	97.29%
Position	73.75	74.50	101.02%
Color Attr	70.00	74.25	106.07%
Average	80.13	81.18	101.31%

Sample Output

MXFP8	BF16 Base Model (Reference)

Generate the Model

auto-round \
    --model_name tencent/HunyuanImage-3.0-Instruct-Distil \
    --model_free \
    --scheme MXFP8 \
    --ignore_layers "vision,guidance_emb,timestep_emb,timestep_r_emb,final_layer,wte" \
    --format auto_round \
    --output_dir HunyuanImage-3.0-Instruct-Distil-MXFP8-AutoRound

Ethical Considerations and Limitations

The model can produce factually incorrect output, and should not be relied on to produce factually accurate information. Because of the limitations of the pretrained model and the finetuning datasets, it is possible that this model could generate lewd, biased or otherwise offensive outputs. Therefore, before deploying any applications of the model, developers should perform safety testing.

Caveats and Recommendations

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. Here are a couple of useful links to learn more about Intel's AI software:

• Intel Neural Compressor
• AutoRound

Disclaimer

The license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model. Please consult an attorney before using this model for commercial purposes.

Cite

@article{cheng2023optimize,
  title={Optimize weight rounding via signed gradient descent for the quantization of llms},
  author={Cheng, Wenhua and Zhang, Weiwei and Shen, Haihao and Cai, Yiyang and He, Xin and Lv, Kaokao and Liu, Yi},
  journal={arXiv preprint arXiv:2309.05516},
  year={2023}
}

arxiv github

Signed-off-by: Lv, Kaokao kaokao.lv@intel.com