Fix bugs shape quantize, outs - Githubissues

open-mmlab / Amphion

Amphion (/æmˈfaɪən/) is a toolkit for Audio, Music, and Speech Generation. Its purpose is to support reproducible research and help junior researchers and engineers get started in the field of audio, music, and speech generation research and development.

https://openhlt.github.io/amphion/

MIT License

4.41k stars 373 forks source link

Fix bugs shape quantize, outs #175

Open huutuongtu opened 4 months ago

huutuongtu commented 4 months ago

Fix size for quantized_buf residual, prosody, phone.

✨ Description

Some audio will have bug with shape when I try voice conversion, so I changed some code to fix:

[1048] outs += out RuntimeError: The size of tensor a (335) must match the size of tensor b (334) at non-singleton dimension 2

👨‍💻 Changes Proposed

[x] Change shape residual, out, x, ...

🧑‍🤝‍🧑 Who Can Review?

@HeCheng0625

✅ Checklist

[ ] Code has been reviewed
[ ] Code complies with the project's code standards and best practices
[ ] Code has passed all tests
[ ] Code does not affect the normal use of existing features
[ ] Code has been commented properly
[ ] Documentation has been updated (if applicable)
[ ] Demo/checkpoint has been attached (if applicable)

yuantuo666 commented 4 months ago

@huutuongtu Hi, please install black and run it to reformat the code, ensuring the GitHub auto-check is successful.

HarryHe11 commented 4 months ago

@HeCheng0625 Hi, yuancheng, could you take a look at this pr about bugs in FaCodec when u r available? The codes look good to me.