[논문] CLAP: Learning Audio Concepts from Natural Language Supervision
https://arxiv.org/abs/2206.04769 CLAP: Learning Audio Concepts From Natural Language SupervisionMainstream Audio Analytics models are trained to learn under the paradigm of one class label to many recordings focusing on one task. Learning under such restricted supervision limits the flexibility of models because they require labeled audio for traininarxiv.org해당 논문을 보고 작성했습니다. Abstract주류 오디오 분석 m..
연구실 공부
2024. 6. 27.