cannot allocate memory

[miracle24]

When I run the att2in2 model with cross entropy loss, I got 'cannot allocate memory' error. I don't know why. The detail is:
File "/usr/lib/python2.7/multiprocessing/process.py", line 130, in start
self._popen = Popen(self)
File "/usr/lib/python2.7/multiprocessing/forking.py", line 121, in init
self.pid = os.fork()
OSError: [Errno 12] Cannot allocate memory

because of my fault, I cannot get all the details. But I known it comes from the detaloader.py which loads the data with multiprocess. It seems that the error didn't happen regularly. Sometimes the whole training can be over without errors, sometimes not. Have you ever met this error? @ruotianluo

[ruotianluo]

It happens to me sometime too, but varies from machine to machine. pytorch/pytorch#1355
I guess it's related to the setting of different machines.

[miracle24]

Thanks a lot. Zan yi ge!