Fix memory leak add throughput tests #20
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
I may remove them later.
Previously, it would always return an empty list. Add support for blank penalty and length penalty, as wenet has them.
Fix GPU memory leak. My pybind11 DLPack integration was incorrect.
Previously, 5 GiB were being created for every test because the librispeech test and dev sets were being loaded every time.
We don't check that RTFx is the same right now. It might be too flakey. Recorded throughput results here: #19
galv
commented
Mar 23, 2023
| openfst_binaries = [f"fst{operation}" for operation in "arcsort closure compile concat connect convert determinize disambiguate encode epsnormalize equal equivalent invert isomorphic map minimize project prune push randgen relabel replace reverse reweight synchronize topsort union".split(" ")] | ||
| fst_binaries = "arpa2fst arpa-to-const-arpa fstdeterminizestar fstrmsymbols fstisstochastic fstminimizeencoded fstmakecontextfst fstmakecontextsyms fstaddsubsequentialloop fstaddselfloops fstrmepslocal fstcomposecontext fsttablecompose fstrand fstdeterminizelog fstphicompose fstcopy fstpushspecial fsts-to-transcripts fsts-project fsts-union fsts-concat transcripts-to-fsts".split(" ") | ||
| fst_binaries.extend(openfst_binaries) | ||
| # ERROR to fix: "fstcompose" is used in mkgraph_ctc.sh... | ||
| fst_binaries = ["arpa2fst", "fsttablecompose", "fstdeterminizestar", "fstminimizeencoded", "fstarcsort", "fstcompile", "fstaddselfloops", "transcripts-to-fsts", "fstconvert"] | ||
| fst_binaries = ["arpa2fst", "fsttablecompose", "fstdeterminizestar", "fstminimizeencoded", "fstarcsort", "fstcompile", "fstaddselfloops", "transcripts-to-fsts", "fstconvert", "fstisstochastic", "fstcompose", "fstrmepslocal"] |
There was a problem hiding this comment.
This is not stricly speaking necessary, but it is helpful. I'm keeping it for now.
| sort $dir/units.txt > $dir/units_sorted.txt | ||
| cmp $dir/units_sorted.txt $dir/units_from_lexicon_sorted.txt | ||
| if [ $? ]; then | ||
| echo "ERROR: Difference in units.txt and units derived from lexicon!" |
There was a problem hiding this comment.
This seems to be a false positive when it triggers. It requires more inspection.
| @@ -383,6 +951,14 @@ def trace_back_stats(r, h, d): | |||
| j = j | |||
| return insertions, substitutions, deletions | |||
|
|
|||
|
|
|||
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.
No description provided.