A prebuilt dataset for OpenAI's task for image-2-latex system. Includes total of ~100k formulas and images splitted into train, validation and test sets. Formulas were parsed from LaTeX sources provided here: http://www. ...