ConformerXXL-PS

Reported on 3 benchmarks across 1 task · 1 paper · 3 SOTA

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

Audio3 results

Speech RecognitiononCHiME-6 eval
Word Error Rate (WER)· 2021-09-27
31
best: 30.7 (ConformerXXL-PS + G-Augment)
SOTA
BigSSL: Exploring the Frontier of Large-Scale Semi-Supervised Learning for Automatic Speech Recognition arXiv:2109.13226
Speech RecognitiononCHiME-6 dev_gss12
Word Error Rate (WER)· 2021-09-27
26.2
best: 26 (ConformerXXL-PS + G-Augment)
SOTA
BigSSL: Exploring the Frontier of Large-Scale Semi-Supervised Learning for Automatic Speech Recognition arXiv:2109.13226
Speech RecognitiononTED-LIUM
Word Error Rate (WER)· 2021-09-27
5
best: 4.6 (Whisper-LLaMa-7b)
SOTA
BigSSL: Exploring the Frontier of Large-Scale Semi-Supervised Learning for Automatic Speech Recognition arXiv:2109.13226