gliclass-multilang-mini-onnx

GLiClass Multilang Mini (ONNX)

This repository provides ONNX-exported versions of knowledgator/gliclass-multilang-mini, a lightweight zero-shot text classification model developed by Knowledgator.

What is ONNX?

ONNX (Open Neural Network Exchange) is an open format designed to represent machine learning models. It enables optimized cross-platform inference using onnxruntime.

Available Variants

• fp32: ~1 GB. Retains 100% of the original model's accuracy. Recommended for production.
• q8: ~350 MB. ~1.2x speedup on CPU, but has significantly worse accuracy than FP32.

Citation

@misc{stepanov2025gliclassgeneralistlightweightmodel,
      title={GLiClass: Generalist Lightweight Model for Sequence Classification Tasks}, 
      author={Ihor Stepanov and Mykhailo Shtopko and Dmytro Vodianytskyi and Oleksandr Lukashov and Alexander Yavorskyi and Mykyta Yaroshenko},
      year={2025},
      eprint={2508.07662},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2508.07662}, 
}