search

easonnie / semanticRetrievalMRS

This is the repo for the paper "Revealing the Importance of Semantic Retrieval for Machine Reading at Scale".
MIT License
59 stars 11 forks source link

system config #5

Open michaelmoju opened 4 years ago

michaelmoju commented 4 years ago

I tried to train the model on a single GTX2080 GPU but got CUDA out of memory error. Can you provide your system configuration? Especially the GPU configuration to successfully train the model. Thanks.

anushkmittal commented 4 years ago

@michaelmoju were you able to figure it out? 15GB GPU doesn't seem to work

michaelmoju commented 4 years ago

I think it needs more than 30G

atish-rnd commented 4 years ago

Facing this issue, tried debugging by deleting model after every epoch and reloading it again from a file. But still runs out of memory in GTX2080 GPU. Can we get the specs for the system it has been trained on previously?