Generic Seq2Seq Decoder LabelScorer #51
Open
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
Remove author line.
curufinwe
requested changes
May 2, 2023
| } | ||
|
|
||
| bool isEqualSequence(const LabelHistoryHandle lhd, const LabelHistoryHandle rhd) const { | ||
| return label_sequence_hash(lhd->labelSeq) == label_sequence_hash(rhd->labelSeq); |
There was a problem hiding this comment.
Comparing hashes for equality is not the same thing as equality. Why is it ok to only check the hashes here?
|
|
||
| // Note: all history have to inherit from LabelHistoryBase | ||
| struct LabelHistoryBase { | ||
| size_t ref_count, cacheHash; |
There was a problem hiding this comment.
Suggested change
| size_t ref_count, cacheHash; | |
| size_t refCount; | |
| size_t cacheHash; |
There was a problem hiding this comment.
Also needs to be changed everywhere else
| }; | ||
|
|
||
| typedef LabelHistoryBase* LabelHistoryHandle; | ||
| typedef std::unordered_map<size_t, LabelHistoryHandle> HistoryCache; |
There was a problem hiding this comment.
size_t here is a manually computed hash and we never test for equality correctly. Ideally we use something like: https://github.com/rwth-i6/rasr/blob/master/src/Lm/NNHistoryManager.hh#L69 for this. And even better: refactor this class and use a common class for this.
| @@ -16,12 +16,300 @@ | |||
| #ifndef TF_LABEL_SCORER_HH | |||
There was a problem hiding this comment.
This file should be inside the Tensorflow module
Co-authored-by: Eugen Beck <curufinwe@users.noreply.github.com>
Co-authored-by: Eugen Beck <curufinwe@users.noreply.github.com>
Co-authored-by: Eugen Beck <curufinwe@users.noreply.github.com>
Co-authored-by: Eugen Beck <curufinwe@users.noreply.github.com>
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.
This PR contains a subset of changes from #45.
These are the changes that introduce the new LabelScorer and TFLabelScorer classes and subclasses.
The code in this PR can be compiled successfully but the new label-scorers are not actually used at this stage.