How to set correctly PYTORCH_CUDA_ALLOC_CONF for stanza? #1521
Closed
ROBERT-MCDOWELL
started this conversation in
General
Replies: 1 comment 2 replies
-
|
I don't know anything about this, but it's going to be even harder to answer without knowing what OS you have or what version of torch you're using |
Beta Was this translation helpful? Give feedback.
2 replies
-
|
sorry it's on Linux Ubuntu 20.04 LTS torch 2.4.1 |
Beta Was this translation helpful? Give feedback.
-
|
ok I think I need a break..... there was a semi colon rather than commas between backend:native etc... |
Beta Was this translation helpful? Give feedback.
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.
-
when I set PYTORCH_CUDA_ALLOC_CONF="backend:native;......"
stanza 1.10.1 on python 3.10 on cuda complains saying it accepts only native and cudaMallocAsync. so I tried
export PYTORCH_CUDA_ALLOC_CONF="backend:cudaMallocAsync" and it fails too.
removing completely PYTORCH_CUDA_ALLOC_CONF let stanza process but it means I cannot set torch anymore....
any idea?
Beta Was this translation helpful? Give feedback.
All reactions