Patch release

v0.16.3

Bump version: 0.16.2 → 0.16.3

Add data redundancy tools

Adds data redundancy tools for duplicate, reverse duplicate and cartesian product relations. See 2020 paper by Akrami et al. for details.

Minor patch and new datasets

• Patch a minor bug in imports
• Add dataset loaders for WN18RR and YAGO3-10

API redefinition and performance improvement

• Redefinition of the models' API (simplified interfaces, renamed LP methods and added get_embeddings method)
• Implementation of the new API for all models
• TorusE implementation fixed
• TransD reimplementation to avoid matmul usage (costly in back-propagation)
• Bilinear models performances improved by avoiding matmul usage
• Addition of feature to negative samplers to generate several negative samples from each fact. Those can be fed directly to
  the models.
• Addition some wrappers for training to utils module.
• Progress bars now make the most of tqdm's possibilities
• Code reformatting
• Docstrings update

Minor patch

• Defined a new class for DataLoader (simpler than the PyTorch one).
• Switched to the use of this class in the various parts of the code.

Minor patch

• Removed the use of torch DataLoader object

Added a method to print results in link prediction evaluator.

v0.15.3

Merge branch 'master' of github.com:torchkge-team/torchkge

Fixed a misfit test.

v0.15.2

Bump version: 0.15.1 → 0.15.2

Cleared the definition of rank in link prediction.

v0.15.1

Merge branch 'master' of github.com:torchkge-team/torchkge

Improved the use of tqdm progress bars.

Merge pull request #128 from torchkge-team/develop

Improved use of tqdm progress bars