-
Notifications
You must be signed in to change notification settings - Fork 13
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Issues with device = "mps"
on Sonoma OS
#144
Comments
Hello @cgoo4 It smells like a silent OOM of the MPS device.
I would recommend you to drastically lower the Let me know if it helps |
Hi @cregouby - I've tried progressively reducing I'm using an M2 Max with 64 GB memory. |
Hello @cgoo4 There are a lot of opened issues currently on device="mps" on sonoma OS (tensorflow, pytorch, ...) so I guess there is nothing we can do until the apple dev fix the MPS code. |
device = "mps"
device = "mps"
on Sonoma OS
Running the example below in a fresh R session,
tabnet_pretrain()
works withdevice = "mps"
, buttabnet_fit()
hangs (no message) and I need to Terminate R to recover. Session info attached.Created on 2024-01-12 with reprex v2.0.2
Session info
The text was updated successfully, but these errors were encountered: