T2R + Pretrain
Reported on 5 benchmarks across 2 tasks · 1 paper · 1 SOTA
Note: results are matched by exact model name. Different papers may use the same name for different model variants.
Natural Language Processing3 results
- BLEU· 2021-03-24SOTA23.8best: 26.5 (StrokeNet)
- BLEU score· 2021-03-2428.7best: 35.14 (Transformer Cycle (Rev))
- BLEU score· 2021-03-2442.1best: 46.4 (Transformer+BT (ADMIN init))
Medical2 results
- Test perplexity· 2021-03-2419.6best: 2.4 (RETRO (7.5B))
- Validation perplexity· 2021-03-2419best: 13.11 (Ensemble of All)