Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Unsupervised Automatic Speech Recognition on LibriSpeech 100 hours (dev-clean)
Loading...
5.2
PER
REBORN
4.636
8.443
12.25
16.057
Feb 6, 2024
PER
WER
Updated 4d ago
Evaluation Results
Method
Method
Links
PER
WER
REBORN
Self-training=HMM ST
2024.02
5.2
9.3
Train from oracle
Boundary=Oracle
2024.02
6.3
12.8
REBORN
2024.02
8.3
12.5
wav2vec-U 2.0
Reproduction=true, Sel...
2024.02
10
15.4
wav2vec-U 2.0
Reproduction=true
2024.02
12.2
17.2
EURO
Backbone=HuBERT
2024.02
15.2
23.1
wav2vec-U
Reproduction=true
2024.02
19.3
20.4
Feedback
Search any
task
Search any
task