TinyBERT-4 14.5M

Reported on 4 benchmarks across 3 tasks · 1 paper

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

Natural Language Processing4 results

Natural Language InferenceonMultiNLI
Matched· 2019-09-23
82.5
best: 92.6 (Turing NLR v5 XXL 5.4B (fine-tuned))
TinyBERT: Distilling BERT for Natural Language Understanding arXiv:1909.10351
Natural Language InferenceonMultiNLI
Mismatched· 2019-09-23
81.8
best: 92.4 (Turing NLR v5 XXL 5.4B (fine-tuned))
TinyBERT: Distilling BERT for Natural Language Understanding arXiv:1909.10351
Semantic Textual SimilarityonSTS Benchmark
Pearson Correlation· 2019-09-23
0.799
best: 0.929 (MT-DNN-SMART)
TinyBERT: Distilling BERT for Natural Language Understanding arXiv:1909.10351
Sentiment AnalysisonSST-2 Binary classification
Accuracy· 2019-09-23
92.6
best: 97.5 (T5-11B)
TinyBERT: Distilling BERT for Natural Language Understanding arXiv:1909.10351