the kernel appears to have died. it will restart automatically.
I try the following:
first: downgrade to cuda 10.2
then try again with jupyter notebook I face the same error and when trying to rerun the cell that cause the error in the forth step I found the following error:
TypeError: Caught TypeError in DataLoader worker process 0
and try to set num_worker to zero. the cell required to redefine some predefined cells but in vain.
second: I implement the code in simple .py file and found the following error:
Illegal instruction (core dumped)
and found the problem in following lines using print statements:
loss.backward()
optimizer.step()
I would recommend to check the system logs to see what’s causing the machine to restart.
A failing Python script should just raise an error, but shouldn’t take down the whole workstation so I guess your current system might encounter some critical issues.
Thanks for your response.
The machine stopped restarting and hanging but the second screen still found.
The error already in the following two lines:
loss.backward()
optimizer.step()
I still think you are facing a potential hardware error as shown in the screenshot so I would recommend to look into the system error logs and try to find any clues what might be wrong.
If you cannot find anything, try to run some RAM tests etc.
I’m beginner. what are system error logs? how could I run some RAM tests?
while install ubuntu 16.04 I faced the second screen above it disappeared after installation.
I faced no original module exists within this kernel while setup gpu driver.