NobleX-AI/vi-stt-mimi-yodas2

YODAS2 Vietnamese: Vietnamese speech as Mimi codes with timed transcripts 6,393.7 hours, 843,598 clips, 73,605,017 words, 30,271 channels of Vietnamese speech from YODAS2 (Li et al.) (YouTube videos under Creative Commons licenses), stored as Kyutai Mimi codes with word-timed transcripts. It is one source of the ~20,000 h on which NobleX-AI trains a streaming Vietnamese speech recognizer (Kyutai's delayed-streams STT, kyutai/stt-1b-en_fr, fine-tuned for Vietnamese). The other… See the full description on the dataset page: https://huggingface.co/datasets/NobleX-AI/vi-stt-mimi-yodas2.

<iframe src="https://silytics.si/embed/item/dc8767b6-4283-4184-bcb3-4afb07ae15c1" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.