3 ms·
Consider looking into language models such as KenLM. It is used by ASR models like wav2letter and DeepSpeech to correct speech-to-text transcripts
by hsson 6y ago
Consider looking into language models such as KenLM. It is used by ASR models like wav2letter and DeepSpeech to correct speech-to-text transcripts