Do we have training examples?

[ChuxiJ]

❓ Questions

Hi encodes team,

great job!

We want to reimplement the results and train a new model in a wider range of the sound dataset and add some denoise/dereverberation functionality. Could you please add some training examples?

[adiyoss]

Hi @ChuxiJ, Thanks for showing interest in our work! Unfortunately, it will take us a bit more time to release the training code, I can not really say when but we will post about it once it is out there.

[adiyoss]

@ChuxiJ, In case you meant something else than the training code, can you please elaborate? :)

[yizhilll]

Hi @adiyoss ,

Though I am not sure what @ChuxiJ was meant, I am also here voting for releasing the codes for training the encodec from scratch. Looking forward to it!

[Cortajarena]

+1 It would be super nice to have this to understand how this architecture is trained from scratch 🚀 Thanks for your work guys!

[keunwoochoi]

+1 training code will be very helpful. thanks!

[dongwon00kim]

+1 I also think the training code is very helpful in understanding the architecture. Thanks.

[rasenganai]

+1 training code would be very helpful. Thanks for this amazing research!

[ZhikangNiu]

+1 training code would be very helpful. Thanks for this amazing research!!!!

[hdmjdp]

+1 training code would be very helpful. Thanks for this amazing research!!!!!!!

[listener17]

Anyone created a training script?

[GitYesm]

+1 training code i need too Thank you !!!!

[zqs01]

+1 training code would be very helpful. Thanks for this amazing research!!!!!!!

[YunyangZeng]

+1 training code would be super helpful!

[thivux]

found a decent implementation with training code here NoFish-528/encodec-pytorch: unofficial implementation of the High Fidelity Neural Audio Compression. hope it helps you guys.

[ZhikangNiu]

found a decent implementation with training code here NoFish-528/encodec-pytorch: unofficial implementation of the High Fidelity Neural Audio Compression. hope it helps you guys.

hhhh, it's my pleasure. And I'm a new comer in audio, maybe there are some bugs. But do not hesitate to contact me if you have any things or questions want to discuss. And I hope we can get together to advances in the field. In my code, I haven't add LM and Balancer, and I only train the 24khz model in LibriTTS 960h.

[DanijelDomazet]

The info about unavailability of a training script should be added to the README.

[leoauri]

+1 Any updates on releasing training code? Would be helpful for related research.. Reproducibility 😅