Speech Recognition on TIMIT

Metric: Percentage error (lower is better)

LeaderboardDataset

Loading chart...

Results

Hide extra data

Sort:

#	Model↕	Percentage error▲	Extra Data	Paper	Date↕	Code
1	wav2vec 2.0	8.3	Yes	wav2vec 2.0: A Framework for Self-Supervised Lea...	2020-06-20	Code
2	vq-wav2vec	11.6	Yes	vq-wav2vec: Self-Supervised Learning of Discrete...	2019-10-12	Code
3	LiGRU + Dropout + BatchNorm + Monophone Reg	14.2	No	The PyTorch-Kaldi Speech Recognition Toolkit	2018-11-19	Code
4	LSTM + Dropout + BatchNorm + Monophone Reg	14.5	No	The PyTorch-Kaldi Speech Recognition Toolkit	2018-11-19	Code
5	wav2vec	14.7	Yes	wav2vec: Unsupervised Pre-training for Speech Re...	2019-04-11	Code
6	GRU + Dropout + BatchNorm + Monophone Reg	14.9	No	The PyTorch-Kaldi Speech Recognition Toolkit	2018-11-19	Code
7	Li-GRU + fMLLR features	14.9	No	Light Gated Recurrent Units for Speech Recognition	2018-03-26	Code
8	RNN + Dropout + BatchNorm + Monophone Reg	15.9	No	The PyTorch-Kaldi Speech Recognition Toolkit	2018-11-19	Code
9	LSTM	16	No	The PyTorch-Kaldi Speech Recognition Toolkit	2018-11-19	Code
10	Li-GRU	16.3	No	The PyTorch-Kaldi Speech Recognition Toolkit	2018-11-19	Code
11	Hierarchical maxout CNN + Dropout	16.5	No	-	-	-
12	RNN	16.5	No	The PyTorch-Kaldi Speech Recognition Toolkit	2018-11-19	Code
13	GRU	16.6	No	The PyTorch-Kaldi Speech Recognition Toolkit	2018-11-19	Code
14	CNN in time and frequency + dropout, 17.6% w/o dropout	16.7	No	-	-	-
15	Light Gated Recurrent Units	16.7	No	Light Gated Recurrent Units for Speech Recognition	2018-03-26	Code
16	RNN-CRF on 24(x3) MFSC	17.3	No	Segmental Recurrent Neural Networks for End-to-e...	2016-03-01	-
17	Bi-RNN + Attention	17.6	No	Attention-Based Models for Speech Recognition	2015-06-24	Code
18	Bi-LSTM + skip connections w/ CTC	17.7	No	Speech Recognition with Deep Recurrent Neural Ne...	2013-03-22	Code
19	QCNN-10L-256FM	19.64	No	Quaternion Convolutional Neural Networks for End...	2018-06-20	Code
20	Soft Monotonic Attention (ours, offline)	20.1	No	Online and Linear-Time Attention by Enforcing Mo...	2017-04-03	Code
21	LAS multitask with indicators sampling	20.4	No	Attention model for articulatory features detect...	2019-07-02	Code
22	LSNN	33.2	No	Long short-term memory and learning-to-learn in ...	2018-03-26	Code

#1wav2vec 2.0SOTA
8.3
Percentage error· Extra Data· 2020-06-20
wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations Code
#2vq-wav2vecSOTA
11.6
Percentage error· Extra Data· 2019-10-12
vq-wav2vec: Self-Supervised Learning of Discrete Speech Representations Code
#3LiGRU + Dropout + BatchNorm + Monophone RegSOTA
14.2
Percentage error· 2018-11-19
The PyTorch-Kaldi Speech Recognition Toolkit Code
#4LSTM + Dropout + BatchNorm + Monophone RegSOTA
14.5
Percentage error· 2018-11-19
The PyTorch-Kaldi Speech Recognition Toolkit Code
#5wav2vec
14.7
Percentage error· Extra Data· 2019-04-11
wav2vec: Unsupervised Pre-training for Speech Recognition Code
#6GRU + Dropout + BatchNorm + Monophone Reg
14.9
Percentage error· 2018-11-19
The PyTorch-Kaldi Speech Recognition Toolkit Code
#7Li-GRU + fMLLR featuresSOTA
14.9
Percentage error· 2018-03-26
Light Gated Recurrent Units for Speech Recognition Code
#8RNN + Dropout + BatchNorm + Monophone Reg
15.9
Percentage error· 2018-11-19
The PyTorch-Kaldi Speech Recognition Toolkit Code
#9LSTM
16
Percentage error· 2018-11-19
The PyTorch-Kaldi Speech Recognition Toolkit Code
#10Li-GRU
16.3
Percentage error· 2018-11-19
The PyTorch-Kaldi Speech Recognition Toolkit Code
#11Hierarchical maxout CNN + Dropout
16.5
Percentage error
No paper
#12RNN
16.5
Percentage error· 2018-11-19
The PyTorch-Kaldi Speech Recognition Toolkit Code
#13GRU
16.6
Percentage error· 2018-11-19
The PyTorch-Kaldi Speech Recognition Toolkit Code
#14CNN in time and frequency + dropout, 17.6% w/o dropout
16.7
Percentage error
No paper
#15Light Gated Recurrent UnitsSOTA
16.7
Percentage error· 2018-03-26
Light Gated Recurrent Units for Speech Recognition Code
#16RNN-CRF on 24(x3) MFSCSOTA
17.3
Percentage error· 2016-03-01
Segmental Recurrent Neural Networks for End-to-end Speech Recognition
#17Bi-RNN + AttentionSOTA
17.6
Percentage error· 2015-06-24
Attention-Based Models for Speech Recognition Code
#18Bi-LSTM + skip connections w/ CTCSOTA
17.7
Percentage error· 2013-03-22
Speech Recognition with Deep Recurrent Neural Networks Code
#19QCNN-10L-256FM
19.64
Percentage error· 2018-06-20
Quaternion Convolutional Neural Networks for End-to-End Automatic Speech Recognition Code
#20Soft Monotonic Attention (ours, offline)
20.1
Percentage error· 2017-04-03
Online and Linear-Time Attention by Enforcing Monotonic Alignments Code
#21LAS multitask with indicators sampling
20.4
Percentage error· 2019-07-02
Attention model for articulatory features detection Code
#22LSNN
33.2
Percentage error· 2018-03-26
Long short-term memory and learning-to-learn in networks of spiking neurons Code