“Gigabye LM”版本间的差异

来自cslt Wiki
跳转至: 导航搜索
3. word-based 3-gram
3. word-based 3-gram
第36行: 第36行:
  
 
== 3. word-based 3-gram ==
 
== 3. word-based 3-gram ==
 +
 +
tri-gram size:
 +
 
{| class="wikitable"
 
{| class="wikitable"
 
|-
 
|-
|      org    ||    th-7  || th-6.5 || th-6
+
|      org    ||    th-7  || th-7/6 || th-6
 
|-
 
|-
 
|10k: 52M  ||      23M ||  8M    ||  4M
 
|10k: 52M  ||      23M ||  8M    ||  4M
 
|-
 
|-
 
|20k: 57M  ||      24M ||  9M    ||  4M
 
|20k: 57M  ||      24M ||  9M    ||  4M
 +
|-
 +
|}
 +
 +
final fst size:
 +
 +
{| class="wikitable"
 +
|-
 +
|      org    ||    th-7        || th-7/6    || th-6
 +
|-
 +
|10k:  -  ||    770M            ||  193M    ||  135M
 +
|-
 +
|20k: -  ||    -            ||  -    || -
 
|-
 
|-
 
|}
 
|}

2012年9月13日 (四) 05:52的版本

1. very initial, without any prunning, character based. Here is the size and perplexity.

The training is with Gigabytes except the cna data, and ppl testing is based on a sub set from the cna data (big52gb applied)

2gram:

25M 2gram.4000.gz: 0 zeroprobs, logprob= -9.39983e+06 ppl= 161.965 ppl1= 177.141


3gram:

47M 3gram.500.gz:0 zeroprobs, logprob= -6.34868e+06 ppl= 85.1361 ppl1= 94.2525

117M 3gram.1000.gz  :0 zeroprobs, logprob= -7.43809e+06 ppl= 80.6408 ppl1= 87.7439

195M 3gram.2000.gz:0 zeroprobs, logprob= -7.95872e+06 ppl= 79.9875 ppl1= 86.5196

221M 3gram.3000.gz:0 zeroprobs, logprob= -8.04799e+06 ppl= 80.2418 ppl1= 86.7277

229M 3gram.4000.gz:0 zeroprobs, logprob= -8.15697e+06 ppl= 82.6585 ppl1= 89.3392

4gram:

205M 4gram.500.gz:0 zeroprobs, logprob= -6.25395e+06 ppl= 79.6739 ppl1= 88.0716

472M 4gram.1000.gz:0 zeroprobs, logprob= -7.21607e+06 ppl= 70.737 ppl1= 76.774


2. pruning the 4k 3gram LM.

Model 1gram 2gram 3gram              size        ppl
1           1e-7  1e-7      1e-7               30M     logprob= -8.55982e+06 ppl= 102.796 ppl1= 111.532
2           1e-6  1e-6      1e-6               5M       logprob= -9.26982e+06 ppl= 150.96   ppl1= 164.9
3           1e-7  1e-6.5    1e-6.5           11M      logprob= -9.09681e+06 ppl= 137.467 ppl1= 149.913


3. word-based 3-gram

tri-gram size:

org th-7 th-7/6 th-6
10k: 52M 23M 8M 4M
20k: 57M 24M 9M 4M

final fst size:

org th-7 th-7/6 th-6
10k: - 770M 193M 135M
20k: - - - -