It's automatic, you don't need to get anything, simply interact with the argparser as usual (i.e. parse_args() and use the results)

I'm working on an existing code-base, the argparse are named opt instead of args, and the auto connection doesn't work.

I want to connect the argparse manually and load the connected argparse If I'm running on a remote machine.

So when you connect it (and you're running remotely), it will automatically fill in the values in the parser itself

How are you calling connect? can you share a snippet of the code?

training_task.connect(opt, name='opt')

And opt is the argparser?

It working, Thanks!

yes

Hi @<1571308010351890432:profile|HurtAnt92> , argparse arguments are automatically connected, you do not need to connect them explicitly. They will be available under the Args section of the hyperparams, using:

task.get_parameters()

Also, how do I know if I'm running remotely or locally ?

Use this:

from clearml.config import running_remotely

if running_remotely():
  print("running remotely, not locally!")

However, note that in most cases you do not need to know if you're running locally or remotely - why do you need to know in this case?