search

ikostrikov / pytorch-a3c

PyTorch implementation of Asynchronous Advantage Actor Critic (A3C) from "Asynchronous Methods for Deep Reinforcement Learning".
MIT License
1.23k stars 279 forks source link

memory overflow #17

Closed dylanthomas closed 7 years ago

dylanthomas commented 7 years ago

I tried to run your code on my ubuntu/Titan-X/Intel i7, but I kept getting memory problems, after about one day. Can you guess why ? I tried both OMP_NUM_THREADS=1 python main.py --env-name "Breakout-v0" --num-processes 8 and OMP_NUM_THREADS=1 python main.py --env-name "Breakout-v0" --num-processes 6 , to see if bringing down num-processes might help. But that did not work out either.

ps.
(from free -m >> mem_usage.txt) total used free shared buff/cache available Mem: 64388 63937 325 14 125 271 Swap: 65451 65451 0 Fri Mar 31 17:55:18 KST 2017

ikostrikov commented 7 years ago

What version of PyTorch are you using?

dylanthomas commented 7 years ago

I just ran 'conda install pytorch torchvision cuda80 -c soumith'

ikostrikov commented 7 years ago

And what version it is? There was a problem with memory overflow in an earlier build of PyTorch.

dylanthomas commented 7 years ago

pytorch 0.1.10 py35_1cu80 [cuda80] soumith (from conda list)

ikostrikov commented 7 years ago

Update PyTorch to the latest version.

dylanthomas commented 7 years ago

ok. many thanks.