How to get a general idea for max volume size on secure cloud?
I have been able to deploy 2TB drives, but what is the standard here? How much storage is there generally per server to estimate what i should expect to be able to get?
20 Replies
Not sure but the hosts requirement should estimate them to around 2tb per gpu recommended
And for 80gbs around 4tb per gpu recommended
That should estimate the avg of machines hosted to runpod should have
Thank you!
Any idea how/if i could work with runpod to get more?
Yep your welcome bro
Hmm like request more storage?
I'd recommend trying to contact support or try network storage
They have bigger limits right
use network storage now
But only available in one region ( usable by machines in one region)
15TB
super slow
Oh super slow?
going to try migrating to nvme volume when needed on pod
can trim down a lot but 4TB might not cut it for some stuff
Can try allocating more and see what happens
If theyre available then great
yep tried that it depends really cant get anything super big though
What do you need the big data storage for btw
computer vision with HD video input
In that case might try working with rp's support maybe
broken into images for faster training
Woo that's cool
You say that until you try to make it fast
Like frame by frame?
Haha yeah everyone getting into technical side won't see this as a fun thing
yes each frame seperate, decoding video takes to long during training
preprocessing took days and I screwed it up multiple times!
dang, you need so much data o_o
do u need it all avaliable every time?
Yea network storage I/O operations has some speed issues with a lot of files coming in and out.
yes for most
realistically i can get it down a lot smaller right now
but later on yes maybe up to 10 terabytes of data will be needed for a train
like right now i bet 5 TB would suffice but I have extra data in network drive