Hi, I’m trying to fine tune LongT5 but it has a higher memory requirement. I was wondering if I should use a larger GPU (via google cloud), if so, which one should I use?
If not, is using deep speed integration enough?
Hi, I’m trying to fine tune LongT5 but it has a higher memory requirement. I was wondering if I should use a larger GPU (via google cloud), if so, which one should I use?
If not, is using deep speed integration enough?