Keyword Spotting on Google Speech Commands

43개 결과 · ⬇ CSV · JSON

Google Speech Commands V1 12

Rank	Model	Google Speech Commands V1 12	Google Speech Commands V2 12	Google Speech Commands V2 2	Google Speech Commands V2 20	Google Speech Commands V2 35	Google Speech Commands V1 2	Google Speech Commands V1 20	Google Speech Commands V1 35	Extra Training Data	Paper	Code	Year
1	TripletLoss-res15	98.56	98.37	–	–	97.0	–	–	–		Learning Efficient Representations for Keyword Spotting with Triplet Loss	roman-vygon/triplet_loss_kws	2021
2	BC-ResNet-8	98.0	98.7	–	–	–	–	–	–		Broadcasted Residual Learning for Efficient Keyword Spotting	google-research/google-research · Qualcomm-AI-research/bcresnet · re9ulus/BC-ResNet · +1	2021
3	Wav2KWS	97.9	98.5	–	97.8	–	–	–	–		Wav2KWS: Transfer Learning from Speech Representations for Keyword Spotting	qute012/Wav2Keyword	2021
4	res 8	97.8	–	–	–	–	–	–	–		Howl: A Deployed, Open-Source Wake Word Detection System	castorini/howl · solyarisoftware/WeBAD	2020
5	KWT-3	97.49 ±0.15	98.56 ±0.07	–	–	97.69 ±0.09	–	–	–		Keyword Transformer: A Self-Attention Model for Keyword Spotting	ARM-software/keyword-transformer · EscVM/EscVM_YT · ID56/Torch-KWT · +7	2021
6	MatchboxNet-3x2x64	97.48	97.63	–	–	–	–	–	–		MatchboxNet: 1D Time-Channel Separable Convolutional Neural Network Architecture for Speech Commands Recognition	google-research/google-research	2020
7	ConvMixer	97.3	98.2	–	–	–	–	–	–		ConvMixer: Feature Interactive Convolution with Curriculum Learning for Small Footprint and Noisy Far-field Keyword Spotting	dianwen-ng/Keyword-Spotting-ConvMixer · IMvision12/keras-vision-models	2022
8	KWT-2	97.27 ±0.08	98.43±0.08	–	–	97.74 ±0.03	–	–	–		Keyword Transformer: A Self-Attention Model for Keyword Spotting	ARM-software/keyword-transformer · EscVM/EscVM_YT · ID56/Torch-KWT · +7	2021
9	KWT-1	97.26±0.18	98.08±0.10	–	–	96.95±0.14	–	–	–		Keyword Transformer: A Self-Attention Model for Keyword Spotting	ARM-software/keyword-transformer · EscVM/EscVM_YT · ID56/Torch-KWT · +7	2021
10	MHAtt-RNN	97.2	98	–	–	–	–	–	–		Streaming keyword spotting on mobile devices	google-research/google-research · qute012/Pytorch-MHAtt-RNN-KWS · Arizona-Voice/blossom	2020
11	NAS1	97.06	–	–	–	–	–	–	–		Neural Architecture Search For Keyword Spotting		2020
12	SSAMBA	96.9	–	–	–	97.4	–	–	–		SSAMBA: Self-Supervised Audio Representation Learning with Mamba State Space Model	siavashshams/ssamba	2024
13	Attention RNN	95.6	96.9	99.4	94.5	93.9	99.2	94.1	94.3		A neural attention model for speech command recognition	google-research/google-research · douglas125/SpeechCmdRecognition · huckiyang/speech_quantum_dl · +5	2018
14	DS-CNN	94.4	–	–	–	–	–	–	–		Hello Edge: Keyword Spotting on Microcontrollers	google-research/google-research · ARM-software/ML-KWS-for-MCU · mindspore-ai/models · +15	2017
15	GRU	93.5	–	–	–	–	–	–	–	✓	Hello Edge: Keyword Spotting on Microcontrollers	google-research/google-research · ARM-software/ML-KWS-for-MCU · mindspore-ai/models · +15	2017
16	LSTM	92.9	–	–	–	–	–	–	–	✓	Hello Edge: Keyword Spotting on Microcontrollers	google-research/google-research · ARM-software/ML-KWS-for-MCU · mindspore-ai/models · +15	2017
17	Basic LSTM	92.0	–	–	–	–	–	–	–		Hello Edge: Keyword Spotting on Microcontrollers	google-research/google-research · ARM-software/ML-KWS-for-MCU · mindspore-ai/models · +15	2017
18	DNN	91.6	–	–	–	–	–	–	–		Hello Edge: Keyword Spotting on Microcontrollers	google-research/google-research · ARM-software/ML-KWS-for-MCU · mindspore-ai/models · +15	2017
19	CNN	84.6	–	–	–	–	–	–	–		Hello Edge: Keyword Spotting on Microcontrollers	google-research/google-research · ARM-software/ML-KWS-for-MCU · mindspore-ai/models · +15	2017
20	WaveFormer	–	98.8	–	–	99.1	–	–	–		Work in Progress: Linear Transformers for TinyML		2024
21	EdgeCRNN 2.0×	–	98.05	–	–	–	–	–	–		EdgeCRNN: an edgecomputing oriented model of acoustic feature enhancement for keyword spotting		2021
22	Embedding + Head	–	97.7	–	–	–	–	–	–		Training Keyword Spotters with Limited and Synthesized Speech Data	dscripka/openwakeword	2020
23	Head without Embedding	–	97.4	–	–	–	–	–	–		Training Keyword Spotters with Limited and Synthesized Speech Data	dscripka/openwakeword	2020
24	TC-ResNet14-1.5	–	96.6	–	–	–	–	–	–		Temporal Convolution for Real-time Keyword Spotting on Mobile Devices	google-research/google-research · hyperconnect/TC-ResNet · Alex-Riviello/KWS_MCU	2019
25	End-to-end KWS model	–	95.55	–	–	–	–	–	–		End-to-end Keyword Spotting using Neural Architecture Search and Quantization		2021
26	MicroNet-KWS-L	–	95.3	–	–	–	–	–	–		MicroNets: Neural Network Architectures for Deploying TinyML Applications on Commodity Microcontrollers	ARM-software/ML-zoo	2020
27	DenseNet-BiLTSM	–	–	–	96.6	–	–	–	–		Effective Combination of DenseNet andBiLSTM for Keyword Spotting		2019
28	LSTM	–	–	–	93.72	–	–	–	–		Multi-layer Attention Mechanism for Speech Keyword Recognition		2019
29	QNN	–	–	–	–	98.60	–	–	–		Towards on-Device Keyword Spotting using Low-Footprint Quaternion Neural Models	DataSenseiAryan/GoogleSpeechCommandLowFootprint	2023
30	M2D	–	–	–	–	98.5	–	–	–		Masked Modeling Duo: Learning Representations by Encouraging Both Networks to Model the Input	nttcslab/m2d	2022
31	EAT-S	–	–	–	–	98.15	–	–	–		End-to-End Audio Strikes Back: Boosting Augmentations Towards An Efficient Audio Classification Network	Alibaba-MIIL/AudioClassfication	2022
32	Audio Spectrogram Transformer	–	–	–	–	98.11	–	–	–		AST: Audio Spectrogram Transformer	YuanGongND/ast · nttcslab/composing-general-audio-repr · pxaris/ccml · +2	2021
33	HTS-AT	–	–	–	–	98.0	–	–	–		HTS-AT: A Hierarchical Token-Semantic Audio Transformer for Sound Classification and Detection	retrocirce/hts-audio-transformer	2022
34	KW-MLP	–	–	–	–	97.56	–	–	–		Attention-Free Keyword Spotting	AI-Research-BD/Keyword-MLP	2021
35	ImportantAug	–	–	–	–	95	–	–	–		ImportantAug: a data augmentation agent for speech	tvanh512/importantAug	2021
36	NAS2	–	–	–	–	–	–	–	–		Neural Architecture Search For Keyword Spotting		2020
37	Quantum CNN	–	–	–	–	–	–	–	–		Decentralizing Feature Extraction with Quantum Convolutional Neural Network for Automatic Speech Recognition	huckiyang/speech_quantum_dl · huckiyang/QuantumSpeech-QCNN	2020
38	TDNN	–	–	–	–	–	–	–	–		Efficient keyword spotting using time delay neural networks		2018
39	PATE-AAE (Differentially-Private)	–	–	–	–	–	–	–	–		PATE-AAE: Incorporating Adversarial Autoencoder into Private Aggregation of Teacher Ensembles for Spoken Command Classification		2021
40	res8 w/ SSN(S=4, A=Sub)	–	–	–	–	–	–	–	–		SubSpectral Normalization for Neural Audio Data Processing		2021
41	res15 w/ SSN(S=4, A=Sub)	–	–	–	–	–	–	–	–		SubSpectral Normalization for Neural Audio Data Processing		2021
42	res15 w/ SSN(S=4, A=Sub) (2019)	–	–	–	–	–	–	–	–		SubSpectral Normalization for Neural Audio Data Processing		2021
43	End-to-End Keyword Spotting on FPGA Usin 자동 추출	–	–	–	–	–	–	–	–		End-to-End Keyword Spotting on FPGA Using Graph Neural Networks with a Neuromorphic Auditory Sensor		2026