Pushing the Limits of Semi-Supervised Learning for Automatic Speech Recognition

Benchmark Model Rank Results
speech-recognition-on-librispeech-test-cleanConformer + Wav2vec 2.0 + SpecAugment-based Noisy Student Training with Libri-Light#28Word Error Rate (WER): 1.4
speech-recognition-on-librispeech-test-otherConformer + Wav2vec 2.0 + SpecAugment-based Noisy Student Training with Libri-Light#14Word Error Rate (WER): 2.6