Mogrifier LSTM + dynamic eval

Reported on 5 benchmarks across 1 task · 1 paper

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Medical5 results

Language ModellingonPenn Treebank (Word Level)
Test perplexity· 2019-09-04
44.9
best: 20.5 (GPT-3 (Zero-Shot))
Mogrifier LSTM arXiv:1909.01792
Language ModellingonPenn Treebank (Word Level)
Validation perplexity· 2019-09-04
44.8
best: 36.1 (BERT-Large-CAS)
Mogrifier LSTM arXiv:1909.01792
Language ModellingonHutter Prize
Bit per Character (BPC)· 2019-09-04
0.988
best: 1.31 (RHN - depth 5 [zilly2016recurrent])
Mogrifier LSTM arXiv:1909.01792
Language ModellingonWikiText-2
Test perplexity· 2019-09-04
38.6
best: 8.21 (SparseGPT (175B, 50% Sparsity))
Mogrifier LSTM arXiv:1909.01792
Language ModellingonWikiText-2
Validation perplexity· 2019-09-04
40.2
best: 15.69 (GPT-2 (fine-tuned))
Mogrifier LSTM arXiv:1909.01792