@<1523703472304689152:profile|UpsetTurkey67> great, thank you! We are taking a look
@<1523701435869433856:profile|SmugDolphin23> let me know if you need any help in reproducing
@<1523701435869433856:profile|SmugDolphin23> it did not help, shall I create smallest example when it does not work and paste it here?
@<1523701435869433856:profile|SmugDolphin23> it took some time, but I was able to cut 90% of the code, just dataloading remains and the problem persists (which is fortunate, as it makes it easy to replicate). Please have a look.
@<1523703472304689152:profile|UpsetTurkey67> It would be great if you could write a script we could use to reproduce
@<1523701087100473344:profile|SuccessfulKoala55> I am using it as follows:
after calling clearml.Task.init()
I create an object:
cache = Cache('/scidata/marek/diskcache')
and then in the loading function I do:
if cache_arg in load_and_crop.cache:
return load_and_crop.cache[cache_arg]
@<1523703472304689152:profile|UpsetTurkey67> can you please open a Github issue as well, so we can better track this one?
@<1523701435869433856:profile|SmugDolphin23> I have checked that when setting auto_connect_frameworks=False it works, but disabling just joblib is not enough.
@<1523703472304689152:profile|UpsetTurkey67> can you expand on how exactly you're using joblib?
but I do agree that some kind of autoconnect may be the issue
@<1523701087100473344:profile|SuccessfulKoala55> FYI
Hi @<1523703472304689152:profile|UpsetTurkey67> ! What if in Task.init
you set auto_connect_frameworks={"joblib": False}
. Do you still have this issue?
@<1523701435869433856:profile|SmugDolphin23> will send later today
@<1523701087100473344:profile|SuccessfulKoala55> I have the same problem with diskcache
I am seeing such warnings clearml.model - WARNING - 9 model found when searching
@<1523701087100473344:profile|SuccessfulKoala55> any ideas what can be the cause?
@<1523701435869433856:profile|SmugDolphin23> None
The problem started appearing when I started to use joblib
with a simple memory caching mechanism.
to avoid loading and cropping a big image