CNN + Bi-RNN + CTC (speech to letters)

Reported on 2 benchmarks across 1 task · 1 paper · 2 SOTA

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

Audio2 results

Speech RecognitiononCHiME real
Percentage error· 2014-12-17
67.94
best: 7.8 (Distortion-Independent + WRBN + Utterance-Wise Recurrent Dropout + Magnitude Features)
SOTA
Deep Speech: Scaling up end-to-end speech recognition arXiv:1412.5567
Speech RecognitiononCHiME clean
Percentage error· 2014-12-17
6.3
best: 3.34 (Deep Speech 2)
SOTA
Deep Speech: Scaling up end-to-end speech recognition arXiv:1412.5567