Training phase question ?

[danionescu0]

Will the train phase, use the "deepspeech_final.pth" model or can it be made to enhance this model ?
I am asking because I just want to improve the model not build one from scratch.

Thanks

[avramandrei]

Hi,

Yes, it is possible. Change the continue_from parameter in deepspeech_pytorch/configs/train_configs.py with the path to deepspeech_final.pth. The model should be initialized with the respective parameters when the training starts.

Let me know if that worked for you.

Bests,
Andrei

[danionescu0]

Thanks, i'll give it a try!

…

On Thu, Apr 15, 2021 at 6:55 PM Andrei-Marius Avram < ***@***.***> wrote: Hi, Yes, it is possible. Change the continue_from parameter in deepspeech_pytorch/configs/train_configs.py with the path to deepspeech_final.pth. The model should be initialized with the respective parameters when the training starts. Let me know if that worked for you. Bests, Andrei — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#2 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABDFS3U7HIDIUAV7SAG35FDTI4EADANCNFSM427U5TVQ> .