---
license: apache-2.0
---
Mol-VL is a Vision-Language Model for Optical Chemical Structure Understanding (OCSU).

To take advantage of existing pretrained VLMs, we adopt the weights from Qwen2-VL. Mol-VL-7B is further finetuned on [Vis-CheBI20](https://huggingface.co/datasets/PharMolix/Vis-CheBI20) training set.

For technical details, please refer to [OCSU](https://arxiv.org/pdf/2501.15415). Training and evaluation scripts will be available recently, stay tuned!

If you find our work useful in your research, please consider citing:

```
@article{fan2025ocsu,
  title={OCSU: Optical Chemical Structure Understanding for Molecule-centric Scientific Discovery},
  author={Fan, Siqi and Xie, Yuguang and Cai, Bowen and Xie, Ailin and Liu, Gaochao and Qiao, Mu and Xing, Jie and Nie, Zaiqing},
  journal={arXiv preprint arXiv:2501.15415},
  year={2025}
}
```