Long-Short Transformer: Efficient Transformers for Language and Vision

Benchmark Model Rank Results
language-modelling-on-enwiki8Transformer-LS (large)#9Bit per Character (BPC): 0.97Number of params: 110M
language-modelling-on-enwiki8Transformer-LS (small)#15Bit per Character (BPC): 0.99
language-modelling-on-text8Transformer-LS (small)#6Bit per Character (BPC): 1.09