search

microsoft / NeuralSpeech

MIT License
1.38k stars 183 forks source link

train new language for videoDubber. #105

Open manmay-nakhashi opened 1 year ago

manmay-nakhashi commented 1 year ago

can we train a new language for videoDubber ? , if you yes is there any documentation or a guide ?.

shivamm7 commented 1 year ago

Is there any update on this issue? Are the code files for the translation_speech_length task and the transformer_speech_length model architecture available?

kdcyberdude commented 1 year ago

Any update on this?

wyh2000 commented 1 year ago

Hi, the files have updated. Please follow https://github.com/microsoft/NeuralSpeech/blob/master/VideoDubber/fairseq/tasks/translation_speech_length.py and https://github.com/microsoft/NeuralSpeech/blob/master/VideoDubber/fairseq/models/transformer_speech_length.py to find more details.

kdcyberdude commented 1 year ago

Thank you, @wyh2000, for sharing the files.

Could you please include the directory and file structure of the dataset in the README for a more comprehensive understanding? Alternatively, if it's more convenient, could you guide us to a script that could assist us in translating the CVSS speech-to-speech dataset into the desired format?

We appreciate your help in advance.