search

steventan0110 / DiffNorm

Official repository for paper: DiffNorm: Self-Supervised Normalization for Non-autoregressive Speech-to-speech Translation
https://arxiv.org/html/2405.13274v1
MIT License
2 stars 1 forks source link

diff_transformer.py is not available inside the directry it referenced #1

Closed mahendraphd closed 2 months ago

mahendraphd commented 3 months ago

from fairseq.models.text_to_speech.diff_transformer import ( ModuleNotFoundError: No module named 'fairseq.models.text_to_speech.diff_transformer'

While trying to run the code getting above module not found error (diff_transformer) form where oi imports two classes DiffusionTransformerModel and base_s2st_transformer_encoder_architecture .

After inspection i found that this diff_transformer.y is not present in the dir, i also checked in fairseq liberary also not present there. Please update this file in the repository or provide any source form where i can ge that python file. Thank you.

steventan0110 commented 2 months ago

@mahendraphd Thanks for spotting the issue. I previously removed that file because it is not actually used, but apparantly the argument processing requires this outdated file. The file is now pushed to the repo and you can find in this commit: https://github.com/steventan0110/DiffNorm/commit/f7e9426147a78132549f5c640cf247cf7e45dcca