Hi,

We are team of students trying to train AI to recognise chem molecules from images. We are seeing good results from our limited data. We are doing this by writing latex code from images and then training the model on images (latex code works as features).

To achieve greater accuracy and coverage, we are in need of a dataset of Hydrocarbons described in Latex (chemfig)/ SMILE format (more than 10K images with latex code).

Please help and point us in the right direction. Is there any openly available dataset in such format.