Hi all,
I have run the first code on Kaggle and have received the following warning (which makes me suspect that think that parallel processing is not allowed on Kaggle):
Code snippet:
learn = vision_learner(dls, resnet34, metrics=error_rate)
learn.fine_tune(1)
Warning (repeated a couple of times during code execution):
/opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock.
self.pid = os.fork()
It has taken ages for the learner to complete the run as opposed to the “few minutes” mentioned in the book.
How can I sort this out and run the learner without this warning, hopefully attaining much faster execution times?
Thanks,
BY