working fine

AgitatedDove14
option 1: use "aiohttp" (async http framework) - https://github.com/aio-libs/aiohttp
option 2: https://stackoverflow.com/questions/5925028/urllib2-post-progress-monitoring
option 3: https://docs.python-requests.org/en/latest/user/advanced/#streaming-uploads + use a generator to yield tiny chunks and print the progress between chunks

no. we are using our own server.

no

yep. they are

nope. nevermind. it worked with the URL approach

and I said - when I upload them directly from URL, it's working.

Thank you, CostlyOstrich36
How can I put this as a feature request?

In the past, we had 2 graphs in each row.

but this extra text doesn't break the URL approach

I logged in to the UI from the local browser. is that enough?

<<KEY>>/experiments/<<KEY>>/info-output/debugImages?columns=selected&columns=type&columns=name&columns=tags&columns=status&columns=project.name&columns=users&columns=started&columns=last_update&columns=last_iteration&columns=parent.name&order=last_update

but now it's not broken.
it only brakes when I try to add different text as the "series" argument.

trains-init: error: unrecognized arguments: config

suffix for URL approach:
/info-output/debugImages?columns=selected&columns=type&columns=name&columns=tags&columns=status&columns=project.name&columns=users&columns=started&columns=last_update&columns=last_iteration&columns=parent.name&order=last_updatesuffix for image approach:
` /info-output/debugImages?columns=selected&columns=type&columns=name&columns=tags&columns=status&columns=project.name&columns=users&columns=started&columns=last_update&columns=last_iteration&columns=parent.name&or...

TRAINS-AGENT setup process
Configuration file already exists: ~/trains.conf
Leaving setup, feel free to edit the configuration file.

the code doesn't work

done

thanks. it seems to work now.

we are still considering whether to upgrade or not.

if it ain't broke, don't fix it

I don't know

jwt.exceptions.DecodeError: It is required that you pass in a value for the "algorithms" argument when calling decode().

TimelyPenguin76 AgitatedDove14
I am using the "report_plotly" method, but it means I have to pass it down to inner scopes/functions and manage it directly, as well as define the plots specifically.
It also eliminates the option to run/debug locally without logging because the code has explicit Logger or Task dependencies, instead of being an external wrapper.
It fills my Experiment board on the server with many debug sessions.

Regarding seaborn - good to know. might be a good alternative ...

and I need it to report results (I wish there was a way to add free text to those debug samples, without using the "title" and "series" args)

but I think i found a clue:
I am using the image title to write details like GT, pred results

I am able to open the images in debug mode just before reporting them

this is part of the URL from clearML server

can you please debug it and explain what was happening? can you please support adding free text to the uploaded image (such as results, predictions, etc.)?

http://<<IP_ADDRESS>>:8081/<<PROJECTNAME>>/<<EXPERIMENT>>/metrics/samples/<<SERIRES_NAME/samples_<<SERIES_NAME>>_00000000.jpeg?X-Amz-Date=1619339287549

I removed ":" "," and "\n" for the free text in "series" and now it works.

I'll try it when a new version comes out. I prefer to use official releases.
for now I'll use just regular characters in this field.
and again - feature request - add free text there.