try to change the files_server in the clearml.conf file to the bucket you want..
for me its help (model + metrics where uploaded to the bucket at the end and not save on the clearml server

its very weird
the train upload to the bucket the files..
ex:
clearml.Task - INFO - Completed model upload to gs://... clearml.Task - INFO - Finished uploadingi have print for the model is uploaded
also i can see all the files in the bucket as i said (model+ metrics )
the only thing that missing is some plots on the clearml server (app ) when i got to the details of the train i cannot see the matrix confusion for example ( but its exists on the bucket )

i thought the error logged it mig...

Hi AgitatedDove14 thanks for the help!
i run it now and in the end the task upload the model for me to the bucket
clearml.Task - INFO - Completed model upload to gs://
but when i check i can see in the bucket only the final model.. do you know how can i save all the logs and all the metric images?

Thanks

yes they are on the clearml-server now

i would like to have it also save on the bucket

1. save space on the clearml server
2. i have the model + all its info in one place on the bucket

i’m using the yolov5 repo
https://github.com/ultralytics/yolov5
its should logged all in the end as I understand

Thanks i can see the files now on the bucket
i saw also error in the end of the training
clearml.storage - ERROR - Failed uploading: HTTPSConnectionPool(host='storage.googleapis.com', port=443): Max retries exceeded with url... (Caused by SSLError(SSLError(1, '[SSL: DECRYPTION_FAILED_OR_BAD_RECORD_MAC] decryption failed or bad record mac
and in the clearml server some images in the PLOTS tab are missing..
is there something else in the conf that i should change ?

again thanks a lot for t...

in the bucket i can see all the files now!
but on the clearml server when i go into the train some of the plots are missing..
like Confusion Matrix ..

for now i think i’m ok
the scalars seems to be right and the metrics there is whats import for me..

the error for uploading is weird

again thanks for the help!!

my gcp instance
try to upgrade openssl.. still got error

Hi Martin
thanks for the quick replay

the use is fixed path for the end model on the bucket after the training
so other process can use it

yes but that’s mean that the app that use the model needs clearml also..

Hi SuccessfulKoala55 thanks for the help
can you point me to the API for creating the credentials?
some examples
Thanks

on the clearml server i can see only :
F1-Confidence Curve, Precision-Confidence Curve, Precision-Recall Curve, Recall-Confidence Curve,but on the bucket all the rest:

API
(in the webui is manually want to generate automatically)

or maybe set it in the conf file of the server like where setting the users and their password

i had that.. in the end the files where uploaded
just some where missing in the clearml server..

yes..
when i change the files_server to be back the clearml server ( save locally ) i can see everything

no..
just the error line i mention

i need to think about that..
anyway thanks a lot

--verbose should print the reporting
what i want is to be able to see some images on the clearml server
same as when i upload the data as --files

yes