Closed AlkaSaliss closed 4 years ago
I also had same issue
This is usually a residual error that comes after a real error in training, unrelated to fastprogress. This is why we always ask for the full stack trace when you file an issue, because it contains more information than the last error message.
Closing this issue, please reopen with a clear reproducer if needed.
When using
learner.fit_flat_cos(n_epochs, cbs=[SaveModelCallback(path)])
it throws the exception However when used without the callback, it works with no errorfast_progress version =
fastprogress-0.2.2
Any idea of what is causing this bug ?