seastar105/emilia-yodas-ko-filtered-pocket-tts

Emilia-YODAS Korean for pocket-tts Train-ready, Qwen3-forced-aligned Korean data derived from seastar105/Emilia-YODAS-KO-filtered. The training transcript is the source JSON text field after NFC and whitespace normalization. pred was not reconciled into the transcript. split utterances hours train 918,609 2276.88 valid 9,472 23.00 The repository includes 196 original TAR shards, portable gzip JSONL manifests with word timestamps, the 4,000-piece Korean… See the full description on the dataset page: https://huggingface.co/datasets/seastar105/emilia-yodas-ko-filtered-pocket-tts.

<iframe src="https://silytics.si/embed/item/9fea4b9c-5dce-4326-9dd9-3a26466b9c58" 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 Information may be incomplete or out of date.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.