AWD-LSTM + dynamic eval

Reported on 4 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.

Medical4 results

Language ModellingonPenn Treebank (Word Level)
Test perplexity· 2017-09-21
51.1
best: 20.5 (GPT-3 (Zero-Shot))
Dynamic Evaluation of Neural Sequence Models arXiv:1709.07432
Language ModellingonPenn Treebank (Word Level)
Validation perplexity· 2017-09-21
51.6
best: 36.1 (BERT-Large-CAS)
Dynamic Evaluation of Neural Sequence Models arXiv:1709.07432
Language ModellingonWikiText-2
Test perplexity· 2017-09-21
44.3
best: 8.21 (SparseGPT (175B, 50% Sparsity))
Dynamic Evaluation of Neural Sequence Models arXiv:1709.07432
Language ModellingonWikiText-2
Validation perplexity· 2017-09-21
46.4
best: 15.69 (GPT-2 (fine-tuned))
Dynamic Evaluation of Neural Sequence Models arXiv:1709.07432