Dataprovider iterator

Add features to dataprovider :

• Pybind iterator for dataProvider with __iter__, __next__ and __len__ implemented functions
• dataProvider shuffle argument to perform random batch permutation for learning
• dataProvider dropLast argument to drop the last non-full batch

Other :

• Add random utils to perform random batch permutation for learning
• Add support for Int8, Int16, Uint8, Uint16 for tensor python binding

changed milestone to %aidge_core - v0.2.0

added LanguagePyBind StatusReview Ready labels

requested review from @cmoineau and @pineapple

assigned to @thibaultallenet

changed the description

added 2 commits

• f43fb81d - Update comments in dataProvider.hpp
• c21cbc8f - Add support for Int8, Int16, Uint8, Uint16 in tensor binding

Compare with previous version

changed the description

The MR is good. These fonctionalities just lack proper tests for now.

You could create a class to set a DataBase from an array of Tensors, then you could write all the required tests.

Apart from that, I approve this MR.

approved this merge request

merged

mentioned in commit 03ce8a78