faster note
Faster Notes with Python and Deep Learning
I will use the pdf slides from this introduction to Reinforcement Learning by David Silver. Let's start by writing some code to convert each slide to the png format using the pdf2image package. To do that we will use the text detector from the ocr.pytorch repository. Follow the instructions there to download the models and save them in the checkpoints folder. Here, we set the input and output folders, then run a loop over all of the input images (the converted pdf slides) and pass then through the single_pic_proc() function that runs the detection and recognition models located in the ocr module.