Distil-Whisper: Robust Knowledge Distillation via Large-Scale Pseudo Labelling
Paper
• 2311.00430 • Published
• 56
ASR
-
-
MSTRE-Net: Multistreaming Acoustic Modeling for Automatic Lyrics Transcription
Paper • 2108.02625 • Published • 1 -
FLAP: Fast Language-Audio Pre-training
Paper • 2311.01615 • Published • 17 -
Audio Flamingo: A Novel Audio Language Model with Few-Shot Learning and Dialogue Abilities
Paper • 2402.01831 • Published • 17
MaskLID: Code-Switching Language Identification through Iterative Masking
Paper • 2406.06263 • Published • 5Note MaskLID方法通過屏蔽主要語言特徵，有效改善了CS場景下的語言識別，特別在多語言混合的句子中表現出色。該方法不僅提高了識別精度，且適用範圍廣泛，能處理大量網絡數據，對未來的自然語言處理應用有重要意義。
-
Interface Design for Self-Supervised Speech Models
Paper • 2406.12209 • Published • 8