提交 a333e64f 编写于 作者: L lifuchen

modified README of transformer_tts and fastspeech

上级 5b804b70
......@@ -129,7 +129,7 @@ python synthesis.py \
--vocoder='griffin-lim' \
```
We currently support two vocoders, ``Griffin-Lim`` algorithm and ``WaveFlow``. You can set ``--vocoder`` to use one of them. If you want to use ``waveflow`` as your vocoder, you need to set ``--config_vocoder`` and ``--checkpoint_vocoder`` which are the path of the config and checkpoint of vocoder. You can download the pre-trained model of ``waveflow`` from [here](https://github.com/PaddlePaddle/Parakeet#vocoders).
We currently support two vocoders, Griffin-Lim algorithm and WaveFlow. You can set ``--vocoder`` to use one of them. If you want to use WaveFlow as your vocoder, you need to set ``--config_vocoder`` and ``--checkpoint_vocoder`` which are the path of the config and checkpoint of vocoder. You can download the pre-trained model of WaveFlow from [here](https://github.com/PaddlePaddle/Parakeet#vocoders).
Or you can run the script file directly.
......
......@@ -98,7 +98,7 @@ python synthesis.py \
--vocoder='griffin-lim' \
```
We currently support two vocoders, ``Griffin-Lim`` algorithm and ``WaveFlow``. You can set ``--vocoder`` to use one of them. If you want to use ``waveflow`` as your vocoder, you need to set ``--config_vocoder`` and ``--checkpoint_vocoder`` which are the path of the config and checkpoint of vocoder. You can download the pre-trained model of ``waveflow`` from [here](https://github.com/PaddlePaddle/Parakeet#vocoders).
We currently support two vocoders, Griffin-Lim algorithm and WaveFlow. You can set ``--vocoder`` to use one of them. If you want to use WaveFlow as your vocoder, you need to set ``--config_vocoder`` and ``--checkpoint_vocoder`` which are the path of the config and checkpoint of vocoder. You can download the pre-trained model of WaveFlow from [here](https://github.com/PaddlePaddle/Parakeet#vocoders).
Or you can run the script file directly.
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册