This is a dataset designed to fine-tune TrOCR on biochemistry terms such as Kcat/Km, various chemical names, and greek letters.
The dataset contains two columns: image which is the path to the image, and label which is the ground-truth text in the image.
image
label