Determining configuration sufficiency for odqa

Hi all,

I am running deep pavlov on a computer that runs ubuntu 19.10. The machine has 32GB of ram. I have a nvidia GPU that has 6GB of ram.
I want to use deep pavlov for open domain question answering. I have my own data in the form of text files. Is my configuration enough for me to train the model on my data? I do not want to train my model from scratch.

I ask because whenever I have tried running deep pavlov on my machine, the train has run and after a bit, died with the only output being “killed” on the terminal screen.

Hi @slimprize,

An ODQA pipeline consists of a ranker and a SQUAD-like model. You can train a ranker on your data and use a pre-trained squad model.

What were your trying to train?