Utterance-Concatenation Curriculum for long-form robust CTC ASR via wav2vec2 fine-tuning on synthetic multi-utterance concatenations
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).