How does Wav2Vec do downstream classification on labeled data? In the paper they use a CTC algorithm for fine-tuning on labeled data, is this how it’s done when someone fine-tunes it on a downstream task as well?
How does Wav2Vec do downstream classification on labeled data? In the paper they use a CTC algorithm for fine-tuning on labeled data, is this how it’s done when someone fine-tunes it on a downstream task as well?