EnCLAP++: Analyzing the EnCLAP Framework for Optimizing Automated Audio Captioning Performance
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914933898215424 |
|---|---|
| author | Kim, Jaeyeon Jeon, Minjeon Jung, Jaeyoon Woo, Sang Hoon Lee, Jinjoo |
| author_facet | Kim, Jaeyeon Jeon, Minjeon Jung, Jaeyoon Woo, Sang Hoon Lee, Jinjoo |
| contents | In this work, we aim to analyze and optimize the EnCLAP framework, a state-of-the-art model in automated audio captioning. We investigate the impact of modifying the acoustic encoder components, explore pretraining with different dataset scales, and study the effectiveness of a reranking scheme. Through extensive experimentation and quantitative analysis of generated captions, we develop EnCLAP++, an enhanced version that significantly surpasses the original. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_01201 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | EnCLAP++: Analyzing the EnCLAP Framework for Optimizing Automated Audio Captioning Performance Kim, Jaeyeon Jeon, Minjeon Jung, Jaeyoon Woo, Sang Hoon Lee, Jinjoo Audio and Speech Processing Artificial Intelligence Sound In this work, we aim to analyze and optimize the EnCLAP framework, a state-of-the-art model in automated audio captioning. We investigate the impact of modifying the acoustic encoder components, explore pretraining with different dataset scales, and study the effectiveness of a reranking scheme. Through extensive experimentation and quantitative analysis of generated captions, we develop EnCLAP++, an enhanced version that significantly surpasses the original. |
| title | EnCLAP++: Analyzing the EnCLAP Framework for Optimizing Automated Audio Captioning Performance |
| topic | Audio and Speech Processing Artificial Intelligence Sound |
| url | https://arxiv.org/abs/2409.01201 |