Training on M1 Pro #851
Unanswered
Cptn-Reynolds
asked this question in
Q&A
Replies: 2 comments
-
Beta Was this translation helpful? Give feedback.
0 replies
-
A fix was merged a few days ago. So I believe you can already get it in the nightly build.
|
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
I can't seem to get the training to work on a Macbook M1 Pro with current macOS 13.4.1. I tried both with virtual env directly in python and via conda. I tried without torchaudio (as stated in the wiki) but also with torch installed (tested both full and nightly).
I can presample, preconfig and prehubert just fine, but then during the training I am stuck on
NotImplementedError: The operator 'aten::_weight_norm_interface' is not currently implemented for the MPS device... As a temporary fix, you can set the environment variable PYTORCH_ENABLE_MPS_FALLBACK= to use the CPU as a fallback for this op.
and get thrown back to the promptNormally I would say well okay I'll have to wait for the implementation, but I know from reddit and other forums, that others got it running on an M1, so what am I missing?
There are several other WARNINGs too
Help would be appreciated :) thanks
Beta Was this translation helpful? Give feedback.
All reactions