Semi-Supervised Speech Recognition via Local Prior Matching

Benchmark Model Rank Results
speech-recognition-on-librispeech-test-cleanLocal Prior Matching (Large Model)#116Word Error Rate (WER): 7.19
speech-recognition-on-librispeech-test-otherLocal Prior Matching (Large Model, ConvLM LM)#113Word Error Rate (WER): 15.28
speech-recognition-on-librispeech-test-otherLocal Prior Matching (Large Model)#115Word Error Rate (WER): 20.84