OpenCodePapers

speech-recognition-on-libri-light-test-other

Speech Recognition
Dataset Link
Results over time
Click legend items to toggle metrics. Hover points for model names.
Leaderboard
PaperCodeWord Error Rate (WER)ABX-acrossABX-withinModelNameReleaseDate
wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations✓ Link5.0wav2vec 2.0 Large-10h-LV-60k2020-06-20
Libri-Light: A Benchmark for ASR with Limited or No Supervision✓ Link56.6TDS 60k pseudo-label + CTC fine-tuning + 4gram-LM2019-12-17
Libri-Light: A Benchmark for ASR with Limited or No Supervision✓ Link69.5CPC unlab-60k+train-10h CPC pretrain + CTC fine-tuning + 4gram-LM2019-12-17
Libri-Light: A Benchmark for ASR with Limited or No Supervision✓ Link13.428.14CPC unlab-60k2019-12-17
Improving Unsupervised Sparsespeech Acoustic Models with Categorical Reparameterization✓ Link20.612.05S6000h-n42-τ2 → 0.12020-05-29