The CMU-AIST submission for the ICME 2025 Audio Encoder Challenge
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918301230170112 |
|---|---|
| author | Bharadwaj, Shikhar Cornell, Samuele Choi, Kwanghee Shim, Hye-jin Deshmukh, Soham Fukayama, Satoru Watanabe, Shinji |
| author_facet | Bharadwaj, Shikhar Cornell, Samuele Choi, Kwanghee Shim, Hye-jin Deshmukh, Soham Fukayama, Satoru Watanabe, Shinji |
| contents | This technical report describes our submission to the ICME 2025 audio encoder challenge. Our submitted system is built on BEATs, a masked speech token prediction based audio encoder. We extend the BEATs model using 74,000 hours of data derived from various speech, music, and sound corpora and scale its architecture upto 300 million parameters. We experiment with speech-heavy and balanced pre-training mixtures to study the impact of different domains on final performance. Our submitted system consists of an ensemble of the Dasheng 1.2 billion model with two custom scaled-up BEATs models trained on the aforementioned pre-training data mixtures. We also propose a simple ensembling technique that retains the best capabilities of constituent models and surpasses both the baseline and Dasheng 1.2B. For open science, we publicly release our trained checkpoints via huggingface at https://huggingface.co/shikhar7ssu/OpenBEATs-ICME-SOUND and https://huggingface.co/shikhar7ssu/OpenBEATs-ICME. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_16273 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | The CMU-AIST submission for the ICME 2025 Audio Encoder Challenge Bharadwaj, Shikhar Cornell, Samuele Choi, Kwanghee Shim, Hye-jin Deshmukh, Soham Fukayama, Satoru Watanabe, Shinji Sound Audio and Speech Processing This technical report describes our submission to the ICME 2025 audio encoder challenge. Our submitted system is built on BEATs, a masked speech token prediction based audio encoder. We extend the BEATs model using 74,000 hours of data derived from various speech, music, and sound corpora and scale its architecture upto 300 million parameters. We experiment with speech-heavy and balanced pre-training mixtures to study the impact of different domains on final performance. Our submitted system consists of an ensemble of the Dasheng 1.2 billion model with two custom scaled-up BEATs models trained on the aforementioned pre-training data mixtures. We also propose a simple ensembling technique that retains the best capabilities of constituent models and surpasses both the baseline and Dasheng 1.2B. For open science, we publicly release our trained checkpoints via huggingface at https://huggingface.co/shikhar7ssu/OpenBEATs-ICME-SOUND and https://huggingface.co/shikhar7ssu/OpenBEATs-ICME. |
| title | The CMU-AIST submission for the ICME 2025 Audio Encoder Challenge |
| topic | Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2601.16273 |