ngqtrung/emilia-dpo
Emilia-DPO DPO preference pairs for English-Mandarin code-switching speech recognition, built by concatenating English and Chinese clips from the Emilia speech corpus (He et al., 2024). These pairs train Audio LLMs with Direct Preference Optimization (DPO) to keep both languages when they transcribe code-switching speech, instead of translating or dropping one language. The paper trains MERaLiON-2-3B, Phi-4-multimodal-instruct and Qwen2-Audio-7B-Instruct on this dataset together… See the full description on the dataset page: https://huggingface.co/datasets/ngqtrung/emilia-dpo.
| Author / organization | ngqtrung |
|---|---|
| License (as declared) | cc-by-nc-4.0 |
| Downloads | 344 (as reported by Hugging Face) |
| Likes | 0 |
| Created | 1 October 2026 |
| Last updated at source | 1 October 2026 |
Popularity, last 3 days
344 now (+33 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Connected in the knowledge graph
Related papers
More from ngqtrung
- ngqtrung/aidm-dogs-vs-cats-dataDataset
- ngqtrung/IMAVBDataset
- ngqtrung/Video-HopChainDataset
- ngqtrung/StreamingBenchDataset
- ngqtrung/mmsuDataset
- ngqtrung/cs-dialogue-dpoDataset
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.