This repository has been archived by the owner on Dec 16, 2022. It is now read-only.
Adding the wikitables parser #1114
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
DeNeutoy
reviewed
Apr 21, 2018
|
|
||
| loss = self._batch_loss(batch, for_training=False) | ||
| val_loss += loss.data.cpu().numpy() | ||
| if loss is not None: | ||
| # You shouldn't necessarily have to compute a loss for validation, so we allow for |
There was a problem hiding this comment.
Actually when is this used?
There was a problem hiding this comment.
This is used by the wikitables parser (which is why it's part of this PR). Our training loss is maximum marginal likelihood, but we don't have a set of candidate logical forms for all of the validation instances. We still can calculate accuracy, though, and we want to include all of the instances in the validation set, to get accurate metrics. So we need the training code to not crash when we're evaluating instances at validation time for which we can't actually compute a loss.
gabrielStanovsky
pushed a commit
to gabrielStanovsky/allennlp
that referenced
this pull request
Sep 7, 2018
* Adding the wikitables parser * Fix dockerfile and docs
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Performance of the parser still doesn't match the original model, but it should just be minor tweaks from here. It'll be a lot easier to manage git if this is merged into master now.
I didn't really change anything in the code, so this is just bringing things over from the
wikitablesbranch, and it has all already been reviewed.There are a few minor changes to
trainer.pythat I don't think got a close enough look previously, and so @DeNeutoy, if you want to give the changes I made there a look, that'd be great.