⚠️ DEPRECATED — please use whisper-chinese-pro

This model has been superseded by:

• 🆕 whisper-chinese-pro (T4, economy)
• ⚡ whisper-chinese-pro-fast (L40S, 3x faster)

The new models include: - ✅ Built-in vocabulary from 教育部辭典 (172k Mandarin + 14k Taiwanese + 14k Hakka words) - ✅ FrequencyWords corpus for en/ja/ko (5k each) - ✅ Word-level timestamps + confidence - ✅ num_speakers control (1-10 or auto) - ✅ 3 input methods (file / URL / Base64) - ✅ Secret-typed hf_token (masked in UI / examples) - ✅ Realtime factor in output - ✅ VTT output format - ✅ Auto language detection

Please migrate to: https://replicate.com/vm6eji6m4/whisper-chinese-pro

(Legacy model. No further updates will be made.)