search

torchkge-team / torchkge

TorchKGE: Knowledge Graph embedding in Python and PyTorch.
Other
381 stars 41 forks source link

ConvKB full implementation #229

Closed lgalarra closed 2 years ago

lgalarra commented 2 years ago

Description

I am trying to train a ConvKB model using an instance of the class Trainer https://github.com/torchkge-team/torchkge/blob/master/torchkge/utils/training.py. However for ConvKB, the call to the method self.model.normalize_parameters() raises a NotImplementedError. Is this necessary for correct training of ConvKB? Is there any other way to train a ConvKB model properly? If not, are there any plans to support it in the future?

Thanks in advance for your support and help!

Best, Luis

armand33 commented 2 years ago

Hi @lgalarra, This seems to be a bug in the code. Looking at the implementation of ConvKB here, the method normalize_parameters raises a NotImplementedError.

Reading again the original article, I think there is no normalization process for the parameters. This method should then just be changed to pass.

I created a new branch and a PR #230 that changes that. As you're working with ConvKB right now, could you confirm there shouldn't be any particular normalization process ? I'll then release the patch.

Thanks

armand33 commented 2 years ago

The PR has been merged into master. This will be included in the next patch release.