Speech Recognition On Libri Light Test Clean
평가 지표
ABX-across
ABX-within
평가 결과
이 벤치마크에서 각 모델의 성능 결과
모델 이름 | ABX-across | ABX-within | Paper Title | Repository |
|---|---|---|---|---|
| S6000h-n42-τ2 → 0.1 | 13.53 | 9.33 | Improving Unsupervised Sparsespeech Acoustic Models with Categorical Reparameterization | |
| CPC unlab-60k | 7.56 | 5.83 | Libri-Light: A Benchmark for ASR with Limited or No Supervision | |
| CPC unlab-60k+train-10h CPC pretrain + CTC fine-tuning + 4gram-LM | - | - | Libri-Light: A Benchmark for ASR with Limited or No Supervision | |
| wav2vec 2.0 Large-10h-LV-60k | - | - | wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations | |
| TDS 60k pseudo-label + CTC fine-tuning + 4gram-LM | - | - | Libri-Light: A Benchmark for ASR with Limited or No Supervision |
0 of 5 row(s) selected.