List your company

Speech Recognition on Librispeech (100 hours) and AISHELL v1

5.4WER (English)FERERO-FT
5.1446.8728.610.328Dec 2, 2024

Evaluation results

MethodLinks
FERERO-FTfine-tuning_mode=FERERO, losses=CTC losses + CPC loss2024.125.44.95.1
LS-FTfine-tuning_mode=Linear Scalarization, losses=CTC losses + CPC loss2024.126.85.96.4
Komatsu et al.2024.127.11——
Init. (M2ASR)model_state=Initial pre-trained2024.127.36.26.7
w/o CPCoptimization_strategy=Sum of supervised CTC losses2024.1211.810.211