Regularizing and Optimizing LSTM Language Models

Open paper
Benchmark Model Rank Results
language-modelling-on-penn-treebank-wordAWD-LSTM + continuous cache pointer#16Test perplexity: 52.8Validation perplexity: 53.9Params: 24M
language-modelling-on-penn-treebank-wordAWD-LSTM#29Test perplexity: 57.3Validation perplexity: 60.0Params: 24M
language-modelling-on-wikitext-2AWD-LSTM + continuous cache pointer#18Test perplexity: 52.0Validation perplexity: 53.8Number of params: 33M
language-modelling-on-wikitext-2AWD-LSTM#30Test perplexity: 65.8Validation perplexity: 68.6Number of params: 33M

Rank counts only results with a code link.