OpenNMT-py error when training with large amount of data

This was useful, thanks! I think the first step should be to turn it off in the first place.

sudo swapoff /swapfile
sudo fallocate -l 25G /swapfile
sudo chmod 600 /swapfile
sudo mkswap /swapfile
sudo swapon /swapfile
sudo swapon --show

In the case one realized it is too much, the file can be deleted and then the process is repeated.

sudo swapoff /swapfile
sudo rm /swapfile

Kind regards,
Yasmin

1 Like

I think this code was changed and I’m having the same issue again.

I made a pull request to apply a fix very similar to last time: