search

fishaudio / fish-speech

Brand new TTS solution
https://speech.fish.audio
Other
13.07k stars 974 forks source link

[BUG]vits和vqgan在自定义数据集上训练后的效果不佳 #248

Closed pangr closed 3 months ago

pangr commented 4 months ago

Feel free to ask any kind of questions in the issues page, but please use English since other users may find your questions valuable.

Describe the bug A clear and concise description of what the bug is.

To Reproduce Steps to reproduce the behavior:

Expected behavior A clear and concise description of what you expected to happen.

Screenshots / log If applicable, add screenshots / logs to help explain your problem.

Additional context Add any other context about the problem here.

我们在自己的数据集上训练了vits和vqgan，但是出来的效果很差，说话人吐字不清，就是有大舌头一样

fake_6000.zip fake_200000.zip 请问是什么问题

数据集是小数据集 只有96条

aixiaodewugege commented 4 months ago

类似的问题

Stardust-minus commented 4 months ago

推荐冻结MRTE或增加辅助训练集

aixiaodewugege commented 4 months ago

推荐冻结MRTE或增加辅助训练集

大佬，有指导如何冻结MRTE的教程吗？辅助训练集是指 别的说话人的数据吗？

leng-yue commented 4 months ago

最好是加点别的说话人, 数据太少训练很容易 model collapse 和灾难性遗忘

pangr commented 4 months ago

最好是加点别的说话人, 数据太少训练很容易 model collapse 和灾难性遗忘

我加了数据集aishell3中的训练集一起训练，并且冻住了MRTE的参数，训练了300000steps，效果还是一样，请问是有什么参数配置需要调整吗