Trellis Network

Reported on 3 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.

Medical3 results

Language ModellingonPenn Treebank (Word Level)
Test perplexity· 2018-10-15
54.19
best: 20.5 (GPT-3 (Zero-Shot))
Trellis Networks for Sequence Modeling arXiv:1810.06682
Language ModellingonPenn Treebank (Character Level)
Bit per Character (BPC)· 2018-10-15
1.158
best: 1.38 (Bipartite Flow)
Trellis Networks for Sequence Modeling arXiv:1810.06682
Language ModellingonWikiText-103
Test perplexity· 2018-10-15
29.19
best: 2.4 (RETRO (7.5B))
Trellis Networks for Sequence Modeling arXiv:1810.06682