3 ms·
Too many reasons to list. Numpy, pandas, matplotlib are all pretty badly designed and very annoying to work with. Dealing with different array types in numpy, p
by kelipso 4y ago
Too many reasons to list. Numpy, pandas, matplotlib are all pretty badly designed and very annoying to work with. Dealing with different array types in numpy, pytorch, etc, I shouldn't have to do. Different array types have different interfaces, they shouldn't, I don't care why. Terrible packaging system. No threading. Slow loops. Etc.