Error in the estimation of NLL in SGVB #85
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
…s, NOT samples drawn from the recognition model) is performed too early: Currently inside the logsumexp, but it should be outside, because one wants to calculate the average nll over the batch, so first one has to estimate the nll for each element in the batch (done in the logsumexp) and the compute the average of these. It can be shown, that the current breze version yields a strictly greater nll than the real estimate.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Summation of the samples within the batch (i.e. different data samples, NOT samples drawn from the recognition model) is performed too early: Currently inside the logsumexp, but it should be outside, because one wants to calculate the average nll over the batch, so first one has to estimate the nll for each element in the batch (done in the logsumexp) and the compute the average of these. It can be shown, that the current breze version yields a strictly greater nll than the real estimate.