TEA-ASR — Taiwan Mandarin ASR 🍵

TEA-ASR (Taiwan Everyday Audio) — Traditional-script + Taiwanese-lexicon ASR with robust Mandarin–English code-switch, adapted from Qwen3-ASR with a tokenizer-first procedure. Output is Traditional Chinese. Set the language hint to Chinese for Taiwan speech (best results).

Format tags (TEA-ASR-1.1 and TEA-ASR-1.1-mini): tick Keep English to transcribe code-switch English verbatim instead of translating it, and pick a numeral style to force Arabic (123) or Chinese (一二三) numbers. First-gen models ignore these.
Model
Language hint
Numeral style (format tag · TEA-ASR-1.1 models)
Examples
Audio (upload or record) Model Language hint Numeral style (format tag · TEA-ASR-1.1 models) Keep English verbatim — don't translate code-switch (format tag · TEA-ASR-1.1 models) Context / hotwords (optional)