30G memory limited

[db12138]

hi. I run the code in a 30G memory RAM but when the code run to
cls.TEXT.build_vocab(train, vectors=vectors) in SICK.iters() using glove.
the program will be killed because of no enough memory.

I want to know is this situation nomal ? how to fix it ? how much memory in needed at less?

[tuzhucheng]

May I know the vocab size of the dataset you are dealing with please? If the vocab size is too large it may cause out of memory.