Browse > Natural Language Processing > Language Modelling > Penn Treebank (Word Level) dataset

Language Modelling on Penn Treebank (Word Level)

Rank Method Test perplexity Validation perplexity Number of params Paper title Year Paper Code
1 35.76 1542M
2019
2 46.54 47.38 22M
2018
3 47.17 48.63 185M
2018
4 47.3 48.0 22M
2018
5 47.69 48.33 22M
2017
6 51.1 51.6 24M
2017
7 52.38 54.12 23M
2018
8 52.8 53.9 24M
2017
9 54.44 56.54 22M
2017
10 54.55 56.72 24M
2019
11 55.3 57.1 24M
2018
12 56.1 58.3 23M
2018
13 56.8 58.9 24M
2017
14 57.3 60.0 24M
2017
15 58.6 60.8 24M
2018
16 64.0 25M
2016
17 65.4 67.9 23M
2016
18 73.2 75.7 24M
2016