Speech Recognition

883 papers with code • 315 benchmarks • 195 datasets

Speech Recognition is the task of converting spoken language into text. It involves recognizing the words spoken in an audio recording and transcribing them into a written format. The goal is to accurately transcribe the speech in real-time or from recorded audio, taking into account factors such as accents, speaking speed, and background noise.

( Image credit: SpecAugment )

Benchmarks

Add a Result

These leaderboards are used to track progress in Speech Recognition

Dataset	Best Model	Compare
LibriSpeech test-clean	Conformer + Wav2vec 2.0 + SpecAugment-based Noisy Student Training with Libri-Light	See all
LibriSpeech test-other	w2v-BERT XXL	See all
TIMIT	wav2vec 2.0	See all
Switchboard + Hub500	IBM (LSTM+Conformer encoder-decoder)	See all
Common Voice German	wav2vec 2.0 XLS-R 1B + TEVR (5-gram)	See all
WSJ eval92	Speechstew 100M	See all
AISHELL-1	MMSpeech With LM	See all
Common Voice Spanish	ConformerCTC-L (4-gram)	See all
Common Voice French	ConformerCTC-L (5-gram)	See all
MediaSpeech	Quartznet	See all
TUDA	Conformer-Transducer (no LM)	See all
SLUE	W2V2-L-LL60K (+ TED-LIUM 3 LM)	See all
swb_hub_500 WER fullSWBCH	IBM (LSTM+Conformer encoder-decoder)	See all
WenetSpeech	Conformer-MoE (64e)	See all
Hub5'00 SwitchBoard	LAS + SpecAugment (with LM, Switchboard mild policy)	See all
Libri-Light test-clean	wav2vec 2.0 Large-10h-LV-60k	See all
Libri-Light test-other	wav2vec 2.0 Large-10h-LV-60k	See all
EasyCom	ReVISE (bf)	See all
WSJ dev93	CTC-CRF ST-NAS	See all
WSJ eval93	Deep Speech 2	See all
Fongbe audio	Triphone (39 features) + LDA and MLLT + SGMM	See all
CHiME-6 dev_gss12	SpeechStew (1B)	See all
Common Voice Italian	Whisper (Large v2)	See all
VIVOS	Vietnamese end-to-end speech recognition using wav2vec 2.0 by VietAI	See all
Common Voice vi	wav2vec2-base-vietnamese-160h (No Language Model)	See all
Europarl-ASR EN Guest-test	mllp_2021_offline_verb	See all
Europarl-ASR EN MEP-test	mllp_2021_offline_filt	See all
Speech Commands	Liquid-S4	See all
LRS3-TED	AV-HuBERT Large	See all
AISHELL-2	Paraformer-large	See all
Switchboard (300hr)	End-to-end LF-MMI	See all
Hub5'00 CallHome	Espresso	See all
Hub5'00 FISHER-SWBD	CTC-CRF	See all
LibriSpeech train-clean-100 test-clean	wav2vec_wav2letter	See all
LibriSpeech train-clean-100 test-other	wav2vec_wav2letter	See all
Common Voice Portuguese	XLSR53 Wav2Vec2 Portuguese by Orlem Santos	See all
Common Voice Frisian	wav2vec2-large-xls-r-1b-frisian	See all
SPGISpeech	Conformer	See all
Common Voice	SpeechStew (1B)	See all
Common Voice English	Whisper (Large v2)	See all
GigaSpeech	Conformer/Transformer-AED	See all
GigaSpeech DEV	Conformer/Transformer-AED	See all
GigaSpeech TEST	Conformer/Transformer-AED	See all
AMI IMH	SpeechStew (100M)	See all
AMI SDM1	SpeechStew (100M)	See all
Switchboard SWBD	SpeechStew (100M)	See all
Switchboard CallHome	SpeechStew (100M)	See all
Tedlium	Whisper	See all
CHiME-6 eval	SpeechStew (1B)	See all
Google Speech Commands - Musan	ImportantAug	See all
LRS2	RAVEn Large	See all
LibriCSS	TS-SEP	See all

Show all 314 benchmarks

Collapse benchmarks

Libraries

Use these libraries to find Speech Recognition models and implementations

espnet/espnet

14 papers

6,638

msalhab96/SpeeQ

13 papers

PaddlePaddle/PaddleSpeech

12 papers

7,613

pytorch/fairseq

11 papers

26,090

See all 19 libraries.

Datasets

Subtasks

Distant Speech Recognition

Sequence-To-Sequence Speech Recognition

Target Speaker Extraction

Accented Speech Recognition

Noisy Speech Recognition

English Conversational Speech Recognition

Latest papers

Most implemented Social Latest No code

DistilXLSR: A Light Weight Cross-Lingual Speech Representation Model

backspacetg/distilxlsr • 2 Jun 2023

Multilingual self-supervised speech representation models have greatly enhanced the speech recognition performance for low-resource languages, and the compression of these huge models has also become a crucial prerequisite for their industrial application.

02 Jun 2023

Paper
Code

Improved DeepFake Detection Using Whisper Features

piotrkawa/deepfake-whisper-features • 2 Jun 2023

With a recent influx of voice generation methods, the threat introduced by audio DeepFake (DF) is ever-increasing.

02 Jun 2023

Paper
Code

SlothSpeech: Denial-of-service Attack Against Speech Recognition Models

0xrutvij/slothspeech • • 1 Jun 2023

We show that popular ASR models like Speech2Text model and Whisper model have dynamic computation based on different inputs, causing dynamic efficiency.

01 Jun 2023

Paper
Code

Perception and Semantic Aware Regularization for Sequential Confidence Calibration

husterpzh/pssr • CVPR 2023

In this work, we find tokens/sequences with high perception and semantic correlations with the target ones contain more correlated and effective information and thus facilitate more effective regularization.

31 May 2023

Paper
Code

Graph Neural Networks for Contextual ASR with the Tree-Constrained Pointer Generator

briansidp/espnet • • 30 May 2023

The incorporation of biasing words obtained through contextual knowledge is of paramount importance in automatic speech recognition (ASR) applications.

30 May 2023

Paper
Code

CommonAccent: Exploring Large Acoustic Pretrained Models for Accent Classification Based on Common Voice

speechbrain/speechbrain • • 29 May 2023

We introduce a simple-to-follow recipe aligned to the SpeechBrain toolkit for accent classification based on Common Voice 7. 0 (English) and Common Voice 11. 0 (Italian, German, and Spanish).

6,059

29 May 2023

Paper
Code

HyperConformer: Multi-head HyperMixer for Efficient Speech Recognition

speechbrain/speechbrain • • 29 May 2023

In particular, multi-head HyperConformer achieves comparable or higher recognition performance while being more efficient than Conformer in terms of inference speed, memory, parameter count, and available training data.

6,059

29 May 2023

Paper
Code