Examples: query, "exact match", wildcard*, wild?ard, wild*rd
Fuzzy search: cake~ (finds cakes, bake)
Term boost: "red velvet"^4, chocolate^2
Field grouping: tags:(+work -"fun-stuff")
Escaping: Escape characters +-&|!(){}[]^"~*?:\ with \, e.g. \+
Range search: properties.timestamp:[1587729413488 TO *] (inclusive), properties.title:{A TO Z}(excluding A and Z)
Combinations: chocolate AND vanilla, chocolate OR vanilla, (chocolate OR vanilla) NOT "vanilla pudding"
Field search: properties.title:"The Title" AND text
Answered
I Want To Clone An Experiment And Then Try With A Different Hyperparameter Value . I Was Wondering If There Is A Way To Explicitly Specify A Hyperparameter. Right Now, I Guess Its Picking It Up From Tensorboard And I Don'T See The Hyperparameters That Are

I want to clone an experiment and then try with a different hyperparameter value . I was wondering if there is a way to explicitly specify a hyperparameter. Right now, I guess its picking it up from tensorboard and I don't see the hyperparameters that are of interest. I don't even see the learning rate . (I am using tf keras)

  
  
Posted 4 years ago
Votes Newest

Answers 5


I only see these tensorflow flags and not any of the actual hyperparameters

  
  
Posted 4 years ago

Hi FriendlyKoala70 , sorry for replying only now...
You are correct - you can use argparse to define hyper parameters that will be automatically logged and visible in the WebApp. You can check out the https://github.com/allegroai/trains/blob/master/examples/hyper_parameters_example.py to see this and other ways you can report and connect hyper-parameters to your experiment.

  
  
Posted 4 years ago

I think I figured it out. I need to use argparse to set the hyperparameters. Then it logs it automatically

  
  
Posted 4 years ago

Thanks .. I tried both the approaches.

  
  
Posted 4 years ago

Or you can do:
param={'key': 123}
task.connect(param)

  
  
Posted 4 years ago
759 Views
5 Answers
4 years ago
one year ago
Tags