Ngram entropy pruning #4594
Merged
Ngram entropy pruning #4594
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
|
Thanks!! Merging. |
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Outdated suggestions cannot be applied.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.
Same (99%) as SRILM.
The unpruned original model (egs/swbd/s5c/data/local/lm/sw1.o3g.kn.gz):
\data
ngram 1=30275
ngram 2=455846
ngram 3=272601
file heldout: 10000 sentences, 118254 words, 0 OOVs
0 zeroprobs, logprob= -250951.4 ppl= 90.50555 ppl1= 132.4765
threshold=4.7e-5:
SRILM:
\data
ngram 1=30275
ngram 2=4681
ngram 3=655
file heldout: 10000 sentences, 118254 words, 0 OOVs
0 zeroprobs, logprob= -290823.7 ppl= 185.1658 ppl1= 287.9481
Our version:
\data
ngram 1=30275
ngram 2=4626
ngram 3=655
file heldout: 10000 sentences, 118254 words, 0 OOVs
0 zeroprobs, logprob= -291397.2 ppl= 187.0819 ppl1= 291.1811
threshold=1e-6:
SRILM:
\data
ngram 1=30275
ngram 2=155789
ngram 3=55781
file heldout: 10000 sentences, 118254 words, 0 OOVs
0 zeroprobs, logprob= -256473.7 ppl= 99.9384 ppl1= 147.5154
Our version:
\data
ngram 1=30275
ngram 2=155465
ngram 3=55781
file heldout: 10000 sentences, 118254 words, 0 OOVs
0 zeroprobs, logprob= -256570.9 ppl= 100.113 ppl1= 147.7948
threshold=3e-8
SRILM:
\data
ngram 1=30275
ngram 2=442951
ngram 3=245963
file heldout: 10000 sentences, 118254 words, 0 OOVs
0 zeroprobs, logprob= -251054.1 ppl= 90.67255 ppl1= 132.7417
Our version:
\data
ngram 1=30275
ngram 2=440476
ngram 3=245963
file heldout: 10000 sentences, 118254 words, 0 OOVs
0 zeroprobs, logprob= -251049.7 ppl= 90.6654 ppl1= 132.7303