Open Randulfe opened 4 years ago
If you want to use the stroke data in NumPy, and work with vector data in Jupyter notebooks, I have processed the raw ndjson's into .npz
format in the dataset directory of this project: https://github.com/dickrnn/dickrnn.github.io
I'm sorry, we don't have a binary version of the dataset. I could not find how to generate them in Google's QuickDraw repository and didn't think it was worth it to further investigate. You should be able to generate .npy files from the simplified images judging by this comment. I hope that helps you!