Adaptive Input Representations for Neural Language Modeling

Benchmark Model Rank Results
language-modelling-on-one-billion-wordAdaptive Input Very Large#7PPL: 23.02Number of params: 1.0BValidation perplexity: 22.92
language-modelling-on-one-billion-wordAdaptive Input Large#11PPL: 23.91Number of params: 0.46BValidation perplexity: 23.83
language-modelling-on-wikitext-103Transformer (Adaptive inputs)#40Test perplexity: 18.70Validation perplexity: 17.97