RoBERTa (ensemble)

Reported on 3 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

Natural Language InferenceonMultiNLI
Mismatched· 2019-07-26
90.2
best: 92.4 (Turing NLR v5 XXL 5.4B (fine-tuned))
SOTA
RoBERTa: A Robustly Optimized BERT Pretraining Approach arXiv:1907.11692
Natural Language InferenceonWNLI
Accuracy· 2019-07-26
89
best: 95.9 (Turing NLR v5 XXL 5.4B (fine-tuned))
RoBERTa: A Robustly Optimized BERT Pretraining Approach arXiv:1907.11692
Sentiment AnalysisonSST-2 Binary classification
Accuracy· 2019-07-26
96.7
best: 97.5 (T5-11B)
RoBERTa: A Robustly Optimized BERT Pretraining Approach arXiv:1907.11692