Barkify

Barkify copied to clipboard

samples

3

could you provide some samples of the results ? thank you

ppisljar

about history_prompt

2

zero shot方式，这个prompt用什么，或者说怎么传入自定义发音人的prompt

hdmjdp

Hi! Do you have any output audio files from your training run to see how results are? Thank you

torphix

作者, 你好: 你的工作对我大有帮助,并且我正在尝试使用你的项目来微调Bark来获取自己的声音、但是我还是有以下困惑需要寻求你的帮助. (1) 我想用自有英语音频数据继续微调Bark text_2.pt 和 coarse_2.pt 请问在生成feats数据时候是否还是沿用 facebook/wav2vec2-large-xlsr-53 和 facebook/wav2vec2-base-960h的组合, 如下图:  期待你的回复、祝好, 谢谢!

jaycase1

Running the traning code asks for semantic_idx. But this component was apparently commented out in process file.

halilergul1

推理的代码里没有加入prompt，如果模型是多发音人训练的，如何指定发音人

hdmjdp

getting this error - cannot pickle '_io.BufferedReader' object' when I try to run train.py for stage 1

asadash123