how to use nsml.load() function properly?

[nkwook]

Hello? our team is trying to use nsml.load() function to use trained model from our code.

We checked load is succeed, but problem occurs when our program tries to use that model.

Is there possible approach to solve this error?

Error code:

Our load code: simclr_model=nsml.load(checkpoint='SimCLR_e200_e60', session='kaist_4/fashion_eval/289')

Saved model in path:

Thanks for your help.

[nsml-admin]

nsml.load() has no return value, so you should not assign the function's return value to a variable.

It is a task to load parameters into the model in the bind_nsml function, so you just call nsml.load() is done

ex)

     if IS_ON_NSML:
         bind_nsml(simclr_model)
         if opts.pause:
             nsml.paused (scope=locals())
     if IS_ON_NSML:
         nsml.load(checkpoint = 'SimCLR_e200_e60', session = 'kaist_4/fashion_eval/289')
     else:
         simclr_model, _, _ = load_model (opts, reload_model = True)

See also : https://n-clair.github.io/ai-docs/_build/html/en_US/contents/load_model.html

[nkwook]

thank u so much!