ngqtrung/cs-dialogue-dpo download history

ngqtrung/cs-dialogue-dpo is an automatic speech recognition dataset on the Hugging Face Hub. In the last 30 days it was downloaded 53 times (53 in the last 7 days), and 53 times in total. It ranks #177,952 among datasets by monthly downloads.

CS-Dialogue-DPO DPO preference pairs for English-Mandarin code-switching speech recognition, built from the spontaneous dialogues of CS-Dialogue (Zhou et al., 2025). These pairs train Audio LLMs with Direct Preference Optimization (DPO) to keep both languages when they transcribe code

Open ngqtrung/cs-dialogue-dpo on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.