List your company

Speech Recognition on Multilingual LibriSpeech (MLS) (test)

0.0322WERgranite-speech-4.0-1b
0.0291280.0498640.07060.091336Dec 6, 2022Jun 22, 2023Jan 7, 2024Jul 23, 2024Feb 7, 2025Aug 24, 2025Mar 11, 2026

Evaluation results

MethodLinks
granite-speech-4.0-1bDecoding Regime=High Accuracy SSD, tau_CTC=0.7, tau_SLM=0.2, Target Language=Spanish2026.030.0322———————779350
granite-speech-4.0-1bDecoding Regime=Full AR, Target Language=Spanish2026.030.0328———————627——
granite-speech-4.0-1bDecoding Regime=High RTFx SSD, tau_CTC=3.0, tau_SLM=0.1, Target Language=Spanish2026.030.0366———————2,490——
granite-speech-4.0-1bDecoding Regime=High Accuracy SSD, tau_CTC=0.7, tau_SLM=0.2, Target Language=German2026.030.0442———————755348
granite-speech-4.0-1bDecoding Regime=High Accuracy SSD, tau_CTC=0.7, tau_SLM=0.2, Target Language=French2026.030.0448———————658139
granite-speech-4.0-1bDecoding Regime=Full AR, Target Language=French2026.030.0451———————634——
granite-speech-4.0-1bDecoding Regime=Full AR, Target Language=German2026.030.0459———————663——
granite-speech-4.0-1bDecoding Regime=High Accuracy SSD, tau_CTC=0.7, tau_SLM=0.2, Target Language=English2026.030.0472———————772142
granite-speech-4.0-1bDecoding Regime=Full AR, Target Language=English2026.030.0474———————702——
granite-speech-4.0-1bDecoding Regime=High RTFx SSD, tau_CTC=3.0, tau_SLM=0.1, Target Language=German2026.030.0491———————2,755——
granite-speech-4.0-1bDecoding Regime=High Accuracy SSD, tau_CTC=0.7, tau_SLM=0.2, Target Language=Average2026.030.0533———————699338
granite-speech-4.0-1bDecoding Regime=Full AR, Target Language=Average2026.030.0547———————629——
granite-speech-4.0-1bDecoding Regime=High RTFx SSD, tau_CTC=3.0, tau_SLM=0.1, Target Language=French2026.030.0552———————2,834——
granite-speech-4.0-1bDecoding Regime=High RTFx SSD, tau_CTC=3.0, tau_SLM=0.1, Target Language=Average2026.030.0573———————2,753——
granite-speech-4.0-1bDecoding Regime=High RTFx SSD, tau_CTC=3.0, tau_SLM=0.1, Target Language=English2026.030.0575———————2,983——
Whispermode=zero-shot2022.120.073——————————
granite-speech-4.0-1bDecoding Regime=High RTFx SSD, tau_CTC=3.0, tau_SLM=0.1, Target Language=Portuguese2026.030.0883———————2,701——
mSLAM-CTCnumber of parameters=2B2022.120.097——————————
granite-speech-4.0-1bDecoding Regime=High Accuracy SSD, tau_CTC=0.7, tau_SLM=0.2, Target Language=Portuguese2026.030.0984———————531810
granite-speech-4.0-1bDecoding Regime=Full AR, Target Language=Portuguese2026.030.1025———————521——
XLS-Rnumber of parameters=1B2022.120.109——————————
REBORN2024.02—0.2090.2690.2820.2470.3990.5150.32———
wav2vec-Usource=Baevski et al. [4]2024.02—0.3250.4020.3980.3330.5810.5980.44———
wav2vec-Usource=Our reproduction2024.02—0.3390.3810.3770.3310.5180.5940.423———
wav2vec-U 2.0source=Liu et al. [30]2024.02—0.2350.3510.3570.2580.4690.4850.359———