Max-We
Thanks for sharing! I'm wondering if you used Docker to train your model in the cloud or did you take another route?
> You've tested this dockerfile, does it need things like ffmpeg for example? For the installation (pip), it does not, however I haven't tried training an entire model with it....
I added a Dockerfile in #128, let me know if this satisfies your requirements.
Thank you, but the system / process memory in use are really low, it seems to be more than enough. Am I missing something?
> Thank you, but the system / process memory in use are really low, it seems to be more than enough. Am I missing something? I was able to resolve...
FYI I played around with this here: https://github.com/Max-We/sf-zero-signal-to-noise
@vvvm23 I will try to implement parts of it this Saturday
FYI I suggested changes for 1. _"rescale the noise schedule to enforce zero terminal SNR"_ 3. _"rescale classifier-free guidance to prevent over-exposure"_ in PR #3664
@patrickvonplaten Thank you for your assistance! The changes look good to me.