List your company
ResearchFormerly WizWand

Automatic Speech Recognition on LibriSpeech 960h (dev-other)

2.4WERw2v-BERT XXL
-1.50424.84851.277.552Aug 7, 2021May 28, 2022Mar 19, 2023Jan 7, 2024Oct 28, 2024Aug 18, 2025Jun 9, 2026
Updated 4mo ago

Evaluation results

MethodLinks
w2v-BERT XXLUnlabeled Data (hrs)=60k, AM Size (B)=1.0, LM Size (B)=0.1, Training Strategy=Pre-training + Self-training, Decoding Strategy=No LM2021.082.4
w2v-BERT XXLUnlabeled Data (hrs)=60k, AM Size (B)=1.0, LM Size (B)=0.1, Training Strategy=Pre-training + Self-training, Decoding Strategy=With LM2021.082.4
HuBERT X-LargeUnlabeled Data (hrs)=60k, AM Size (B)=1.0, Training Strategy=Pre-training Only, Decoding Strategy=With LM2021.082.5
w2v-BERT XXLUnlabeled Data (hrs)=60k, AM Size (B)=1.0, LM Size (B)=0.1, Training Strategy=Pre-training Only, Decoding Strategy=With LM2021.082.6
w2v-Conformer XXLUnlabeled Data (hrs)=60k, AM Size (B)=1.0, LM Size (B)=0.1, Training Strategy=Pre-training + Self-training, Decoding Strategy=With LM2021.082.6
w2v-Conformer XXL+Unlabeled Data (hrs)=60k, AM Size (B)=1.1, LM Size (B)=0.1, Training Strategy=Pre-training + Self-training, Decoding Strategy=With LM2021.082.6
w2v-BERT XLUnlabeled Data (hrs)=60k, AM Size (B)=0.6, LM Size (B)=0.1, Training Strategy=Pre-training + Self-training, Decoding Strategy=No LM2021.082.6
w2v-BERT XLUnlabeled Data (hrs)=60k, AM Size (B)=0.6, LM Size (B)=0.1, Training Strategy=Pre-training + Self-training, Decoding Strategy=With LM2021.082.6
w2v-BERT XXLUnlabeled Data (hrs)=60k, AM Size (B)=1.0, LM Size (B)=0.1, Training Strategy=Pre-training Only, Decoding Strategy=No LM2021.082.7
wav2vec 2.0Unlabeled Data (hrs)=60k, AM Size (B)=0.3, LM Size (B)=> 0.4, Training Strategy=Pre-training + Self-training, Decoding Strategy=With LM2021.082.7
w2v-Conformer XXLUnlabeled Data (hrs)=60k, AM Size (B)=1.0, LM Size (B)=0.1, Training Strategy=Pre-training + Self-training, Decoding Strategy=No LM2021.082.7
w2v-Conformer XXL+Unlabeled Data (hrs)=60k, AM Size (B)=1.1, LM Size (B)=0.1, Training Strategy=Pre-training + Self-training, Decoding Strategy=No LM2021.082.7
w2v-BERT XLUnlabeled Data (hrs)=60k, AM Size (B)=0.6, LM Size (B)=0.1, Training Strategy=Pre-training Only, Decoding Strategy=With LM2021.082.8
w2v-BERT XLUnlabeled Data (hrs)=60k, AM Size (B)=0.6, LM Size (B)=0.1, Training Strategy=Pre-training Only, Decoding Strategy=No LM2021.082.9
wav2vec 2.0Unlabeled Data (hrs)=60k, AM Size (B)=0.3, LM Size (B)=> 0.4+, Training Strategy=Pre-training Only, Decoding Strategy=With LM2021.083
HuBERT LargeUnlabeled Data (hrs)=60k, AM Size (B)=0.3, Training Strategy=Pre-training Only, Decoding Strategy=With LM2021.083
w2v-Conformer XXLUnlabeled Data (hrs)=60k, AM Size (B)=1.0, LM Size (B)=0.1, Training Strategy=Pre-training Only, Decoding Strategy=With LM2021.083
Conformer L with NSTUnlabeled Data (hrs)=60k, AM Size (B)=0.1, LM Size (B)=0.1, Training Strategy=Self-training Only, Decoding Strategy=With LM2021.083.1
wav2vec 2.0Unlabeled Data (hrs)=60k, AM Size (B)=0.3, LM Size (B)=> 0.4, Training Strategy=Pre-training + Self-training, Decoding Strategy=No LM2021.083.1
w2v-Conformer XLUnlabeled Data (hrs)=60k, AM Size (B)=0.6, LM Size (B)=0.1, Training Strategy=Pre-training Only, Decoding Strategy=With LM2021.083.2
w2v-Conformer XXLUnlabeled Data (hrs)=60k, AM Size (B)=1.0, LM Size (B)=0.1, Training Strategy=Pre-training Only, Decoding Strategy=No LM2021.083.2
Conformer L with NSTUnlabeled Data (hrs)=60k, AM Size (B)=0.1, LM Size (B)=0.1, Training Strategy=Self-training Only, Decoding Strategy=No LM2021.083.3
w2v-Conformer XLUnlabeled Data (hrs)=60k, AM Size (B)=0.6, LM Size (B)=0.1, Training Strategy=Pre-training Only, Decoding Strategy=No LM2021.083.5
HuBERT-L + Llama 8B (4bit, r=32)Train=Joint, Decode=PF-S2T2026.063.6
Whisper-LDecode=Beam2026.063.9
HuBERT-LTrain=CTC, Decode=Greedy2026.064
HuBERT-L + Llama 8B (4bit, r=32)Train=PF-S2T, Decode=PF-S2T2026.064.2
HuBERT-L + Llama 8B (4bit, r=32)Train=Joint, Decode=S2T2026.064.2
Conformer LUnlabeled Data (hrs)=N/A, AM Size (B)=0.1, LM Size (B)=0.1, Training Strategy=Trained from Scratch, Decoding Strategy=No LM2021.084.4
wav2vec 2.0Unlabeled Data (hrs)=60k, AM Size (B)=0.3, LM Size (B)=> 0.4+, Training Strategy=Pre-training Only, Decoding Strategy=No LM2021.084.5
Whisper-L + Llama 8B (4bit, r=32)Train=Joint, Decode=PF-S2T2026.064.9
HuBERT-L + Llama 8B (4bit, r=32)Train=S2T, Decode=S2T2026.064.9
Whisper-L + Llama 8B (4bit, r=32)Train=Joint, Decode=S2T2026.065
Whisper-L + Llama 8B (4bit, r=32)Train=PF-S2T, Decode=PF-S2T2026.065.5
Whisper-L + Llama 8B (4bit, r=32)Train=S2T, Decode=S2T2026.065.7
PDS-Base-16Architecture=Conformer, Group=E, L=12, dh=512, dff=2048, h=8, #Params=114M2023.056.31
Stack-4 (Baseline)Architecture=Conformer, Group=E, L=12, dh=512, dff=2048, h=8, #Params=113M2023.056.47
PDS-Base-8Architecture=Conformer, Group=E, L=12, dh=512, dff=2048, h=8, #Params=113M2023.056.54
Stack-4 (wenet)Architecture=Conformer, Group=E, L=12, dh=512, dff=2048, h=8, #Params=109M2023.056.6
PDS-Base-32Architecture=Conformer, Group=E, L=12, dh=512, dff=2048, h=8, #Params=119M2023.056.67
PDS-Base-16Architecture=Conformer, Group=D, L=12, dh=256, dff=2048, h=4, #Params=46M2023.056.97
PDS-Base-8Architecture=Conformer, Group=D, L=12, dh=256, dff=2048, h=4, #Params=46M2023.057.07
PDS-Base-32Architecture=Conformer, Group=D, L=12, dh=256, dff=2048, h=4, #Params=47M2023.057.15
Stack-4 (Baseline)Architecture=Conformer, Group=D, L=12, dh=256, dff=2048, h=4, #Params=45M2023.057.23
PDS-Base-8Architecture=Transformer, Group=B, L=12, dh=512, dff=2048, h=8, #Params=75M2023.057.46
AIPAArchitecture=Enc-CTC, Data Augmentation=SpecAugment + AIPA, Auxiliary Loss=InterCTC, Technique=COS2024.067.61
PDS-Base-16Architecture=Transformer, Group=B, L=12, dh=512, dff=2048, h=8, #Params=76M2023.057.73
PDS-Deep-32Architecture=Transformer, Group=C, L=30, dh=256, dff=2048, h=4, #Params=55M2023.057.77
PDS-Deep-16Architecture=Transformer, Group=C, L=30, dh=256, dff=2048, h=4, #Params=54M2023.057.83
PDS-Deep-8Architecture=Transformer, Group=C, L=30, dh=256, dff=2048, h=4, #Params=53M2023.057.9
PDS-Base-32Architecture=Transformer, Group=B, L=12, dh=512, dff=2048, h=8, #Params=82M2023.057.94
Stack-4 (fairseq)Architecture=Transformer, Group=B, L=12, dh=512, dff=2048, h=8, #Params=71M2023.058
Stack-4 (Baseline)Architecture=Transformer, Group=B, L=12, dh=512, dff=2048, h=8, #Params=71M2023.058.15
Stack-4 (Baseline)Architecture=Transformer, Group=C, L=30, dh=256, dff=2048, h=4, #Params=53M2023.058.51
PDS-Base-8Architecture=Transformer, Group=A, L=12, dh=256, dff=2048, h=4, #Params=30M2023.058.63
PDS-Base-16Architecture=Transformer, Group=A, L=12, dh=256, dff=2048, h=4, #Params=30M2023.058.73
Stack-4 (fairseq)Architecture=Transformer, Group=A, L=12, dh=256, dff=2048, h=4, #Params=30M2023.058.9
MELDSize=260M, Hrs=960, Beam search size=52026.059
Stack-4 (Baseline)Architecture=Transformer, Group=A, L=12, dh=256, dff=2048, h=4, #Params=30M2023.059.26
PDS-Base-32Architecture=Transformer, Group=A, L=12, dh=256, dff=2048, h=4, #Params=31M2023.059.31
BaselineArchitecture=Enc-CTC, Data Augmentation=SpecAugment, Auxiliary Loss=InterCTC2024.069.34
MELDSize=200M, Hrs=960, Beam search size=52026.059.8
dMel (ASR)Size=258M, Hrs=960, Beam search size=52026.0510.3
MELD w/o SpecAugSize=200M, Hrs=960, Beam search size=52026.0512.5
Codec-LMSize=200M, Hrs=960, Beam search size=52026.0516.5
Codec-LM w/o codebook initSize=200M, Hrs=960, Beam search size=52026.05100