Sigh, cloud computing still difficult
. Put up a GPU instance for a remote intern on GCP. Needs to do lots of i/o, so add NVME disk. Server gets terminated after a week or so -- expected. But now the disk is wiped and a bunch of state is lost on a short internship.
-
-
or wrote snapshots to the blob store
-
I often mount the blob store as a disk with gcsfuse , which works ok. Writing blob-specific code is a hassle, especially since it probably runs differently on your laptop
Kraj razgovora
Novi razgovor -
-
-
This sort of thing teaches people to keep reproducibility in mind from the very first line of code. Reproducible code doesn't need intermediate state, only the source dataset (ideally).
-
I mean yeah but this is a bit "let them eat cake" , no? It's reasonable for a quick internship project to prioritise iteration speed
- Još 1 odgovor
Novi razgovor -
-
-
Periodic snapshot of the attached disk?
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
-
-
i guess this is why companies hire dev ops and data engineers as separate jobs. tough for a small place.
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
-
-
Maybe
@HelloPaperspace could be an easier alternative?Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.
Author of the
Founder