EnCLAP: Combining Neural Audio Codec and Audio-Text Joint Embedding for Automated Audio Captioning
We propose EnCLAP, a novel framework for automated audio captioning. EnCLAP employs two acoustic representation models, EnCodec and CLAP, along with a pretrained language model, BART. We also introduce a new training objective called masked codec modeling that improves acoustic awareness of the pretrained language model. Experimental results on AudioCaps and Clotho demonstrate that our model surpasses the performance of baseline models. Source code will be available at https://github.com/jaeyeonkim99/EnCLAP . An online demo is available at https://huggingface.co/spaces/enclap-team/enclap .
PDF AbstractCode
Spaces
AudioCaps
Clotho
| Task | Dataset | Model | Metric Name | Metric Value | Global Rank | Benchmark |
|---|---|---|---|---|---|---|
| Audio captioning | AudioCaps | EnCLAP-large | CIDEr | 0.8029 | # 2 | |
| SPIDEr | 0.4954 | # 1 | ||||
| SPICE | 0.1879 | # 1 | ||||
| METEOR | 0.2554 | # 1 | ||||
| Audio captioning | AudioCaps | EnCLAP-base | CIDEr | 0.7795 | # 4 | |
| SPIDEr | 0.4829 | # 3 | ||||
| SPICE | 0.1863 | # 2 | ||||
| METEOR | 0.2473 | # 3 |
