Share your thoughts, 1 month free Claude Pro on usSee more

Speech Recognition on Hub5'00 CH (test)

12.2WER

n-gram + model M + NNLM

Updated 1mo ago

Evaluation Results

Method	Links
n-gram + model M + NNLM 2016.04		12.2
n-gram + NNLM 2016.04		12.4
n-gram + model M 2016.04		12.6
85K vocab n-gram baseline 2016.04		13.7
30K vocab n-gram baseline 2016.04		14.4
hybrid 2018.05		17.3
hybrid 2018.05		19
Deep Speech 2014.12		19.3
LF MMI 2018.05		19.3
DNN-HMM 2014.12		23.7
DNN-HMM sMBR 2014.12		24.1
CTC 2018.05		25.3
attention model 2018.05		25.7
attention model 2018.05		26.1
DNN-HMM 2014.12		26.3
attention model 2018.05		27.1
Deep Speech 2014.12		31.8
CTC 2018.05		31.8
CTC 2018.05		32.1
GMM-HMM BMMI 2014.12		33
CTC 2018.05		37.1
CTC 2018.05		40.2
attention 2018.05		40.8
attention 2018.05		46
attention 2018.05		48.2
attention 2018.05		50.4
attention 2018.05		52.7
CTC 2018.05		56.1