Gradual Learning of Recurrent Neural Networks

Benchmark Model Rank Results
language-modelling-on-penn-treebank-wordGL-LWGC + AWD-MoS-LSTM + dynamic eval#6Test perplexity: 46.34Validation perplexity: 46.64Params: 26M
language-modelling-on-wikitext-2GL-LWGC + AWD-MoS-LSTM + dynamic eval#14Test perplexity: 40.46Validation perplexity: 42.19