LingweiMeng / Whisper-Sidecar

The implementation for "Empowering Whisper as a Joint Multi-Talker and Target-Talker Speech Recognition System".
17 stars 0 forks source link

能提供一个样例吗 #1

Open weiweisss opened 2 months ago

weiweisss commented 2 months ago

能提供一个样例来说明如何使用该模型,识别多说话人音频中指定说话人说的话吗?

LingweiMeng commented 2 months ago

Hi. I've update the README.md for clarity.

For target-talker ASR training and reference, please first organize your data according to what is mentioned in the README. Then use python finetune.py --taget_asr True to train your model, and finally use python evaluation.py --taget_asr True for reference.