Speech Recognition On Timit

评估指标

Percentage error

评测结果

各个模型在此基准测试上的表现结果

		Paper Title	Repository
LSNN	33.2	Long short-term memory and learning-to-learn in networks of spiking neurons
LAS multitask with indicators sampling	20.4	Attention model for articulatory features detection
Soft Monotonic Attention (ours, offline)	20.1	Online and Linear-Time Attention by Enforcing Monotonic Alignments
QCNN-10L-256FM	19.64	Quaternion Convolutional Neural Networks for End-to-End Automatic Speech Recognition
Bi-LSTM + skip connections w/ CTC	17.7	Speech Recognition with Deep Recurrent Neural Networks
Bi-RNN + Attention	17.6	Attention-Based Models for Speech Recognition
RNN-CRF on 24(x3) MFSC	17.3	Segmental Recurrent Neural Networks for End-to-end Speech Recognition	-
Light Gated Recurrent Units	16.7	Light Gated Recurrent Units for Speech Recognition
CNN in time and frequency + dropout, 17.6% w/o dropout	16.7	-	-
GRU	16.6	The PyTorch-Kaldi Speech Recognition Toolkit
Hierarchical maxout CNN + Dropout	16.5	-	-
RNN	16.5	The PyTorch-Kaldi Speech Recognition Toolkit
Li-GRU	16.3	The PyTorch-Kaldi Speech Recognition Toolkit
LSTM	16.0	The PyTorch-Kaldi Speech Recognition Toolkit
RNN + Dropout + BatchNorm + Monophone Reg	15.9	The PyTorch-Kaldi Speech Recognition Toolkit
GRU + Dropout + BatchNorm + Monophone Reg	14.9	The PyTorch-Kaldi Speech Recognition Toolkit
Li-GRU + fMLLR features	14.9	Light Gated Recurrent Units for Speech Recognition
wav2vec	14.7	wav2vec: Unsupervised Pre-training for Speech Recognition
LSTM + Dropout + BatchNorm + Monophone Reg	14.5	The PyTorch-Kaldi Speech Recognition Toolkit
LiGRU + Dropout + BatchNorm + Monophone Reg	14.2	The PyTorch-Kaldi Speech Recognition Toolkit

0 of 22 row(s) selected.