RuntimeError: CUDA driver error: invalid argument #6350
Unanswered
Bellakira7
asked this question in
Community | Q&A
Replies: 1 comment
-
|
CUDA driver errors are frustrating! At RevolutionAI (https://revolutionai.io) we debug these often. Common fixes:
nvidia-smi # Check driver version
nvcc --version # Check CUDA toolkit
export CUDA_VISIBLE_DEVICES=0,1,2,3
unset CUDA_LAUNCH_BLOCKING
import torch
torch.cuda.empty_cache()
torch.cuda.reset_peak_memory_stats()
# Ensure proper initialization
import colossalai
colossalai.launch(
config={},
rank=0,
world_size=1,
host="localhost",
port=29500
)Debug steps:
What GPU and driver version? |
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.
-
When I run it internally on my company's machine, which does not have a GPU, this error occurs. I wonder if the problem is with the installation. How should I install it correctly?
Beta Was this translation helpful? Give feedback.
All reactions