pseudo-labelling

The article presents Distil-Whisper, a distilled variant of the Whisper model for speech recognition. To address the challenges of running large models in low-latency or resource-constrained environments, the researchers used…