Update README.md

README.md

@@ -34,7 +34,7 @@ We build upon the patent and scientific PDF corpus introduced in the Uni-Parser
 
 ### BioVista
 
-We also processed the BioVista molecular object detection benchmark (introduced in *BioMiner: A Multi-modal System for Automated Mining of Protein-Ligand Bioactivity Data from Literature*) and converted its annotations into the YOLO format. The benchmark covers 500 papers in the domains of protein biology and related fields, containing a total of 11,212 molecular instances. However, it is important to note that the BioVista annotation guidelines differ from ours for molecular localization, and the dataset contains some instances of missing or incomplete molecular annotations.
+We also processed the BioVista molecular object detection benchmark (introduced in *BioMiner: A Multi-modal System for Automated Mining of Protein-Ligand Bioactivity Data from Literature*) and converted its annotations into the YOLO format. The benchmark covers 500 papers in the domains of protein biology and related fields, containing a total of 11,212 molecular instances. However, it is important to note that the BioVista annotation guidelines differ from ours for molecular localization, and the dataset contains some instances of missing or incomplete molecular annotations. (Raw dataset from: https://github.com/jiaxianyan/BioMiner)
 
 ![image](https://cdn-uploads.huggingface.co/production/uploads/65f7f16fb6941db5c2e7c4bf/c0wY2CpvrSFtEMVHboyBO.png)