Implemented random triplet batching

Handling #43 (closed)

I've implemented 2 steps out 3.

The triplet selection I'll implement later; #44 (closed) is the priority now.

Do you have some time to review it @amohammadi ?

Thanks

added 1 commit

• 83bce596 - Created an estimator for triplet networks

Compare with previous version

added 2 commits

• eea4aa58 - changed the triplet_loss to triple_loss_deprecated
• 046ac8d8 - Appended the keyword as a feature

Compare with previous version

added 1 commit

• 9c5df446 - Solving python3 issue

Compare with previous version

unmarked as a Work In Progress

changed the description

Can I merge this @amohammadi ?

Thanks

@tiago.pereira except for the many options that you have in a couple of your functions, it looks good. I have some ideas on how to simplify this using bob.bio.base!102 (merged) but that is not merged yet.

Thank you for putting the effort of refactoring this package. Please go ahead and merge it.

I agree, there are a lot of options. We need to rethink this. I think an OO approch would be cleaner (with aggregation).

merged

mentioned in commit aeaf3303