Self-supervised wav2vec2-style pretraining with consistency-regularized CTC fine-tuning for noise-robust ASR
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).