Thanks for sharing. I'm curious about a couple of things: - What's your favourite way to preprocess images? Most real world data isn't already segmented into i

Thanks for sharing. I'm curious about a couple of things: - What's your favourite way to preprocess images? Most real world data isn't already segmented into images of complete words, and that preprocessing step could affect the quality of the final output. - Do you know of any open source models I could use as a starting point to recognise text chunks with some 2D structure, like math equations that span multiple lines?