Paperspace GPU+ machine run out of memory?


#1

Hi,

I configured a GPU+ machine to run my machine learning python code on Jupyter Notebook. But a memory issue occurred, my code crashed due to that memory cannot be allocated.

FYI, this is the entire error message:
OSError: [Errno 12] Cannot allocate memory.

So, how can I check the amount of memory used when code occurred an error? Does anybody have solution to solve this?

Thanks,
Jiaming


#2

Hi Jiaming, you could try typing !nvidia-smi to view the GPU utilization and performance.