Hierarchical Multiscale Recurrent Neural Networks

Open paper
Benchmark Model Rank Results
language-modelling-on-enwiki8LN HM-LSTM#36Bit per Character (BPC): 1.32Number of params: 35M
language-modelling-on-text8LayerNorm HM-LSTM#18Bit per Character (BPC): 1.29Number of params: 35M

Rank counts only results with a code link.