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
Hello, I Have A Question Regarding Clearml Hyperparameteroptimizer I Have A Model That Has As Input

Hello, I have a question regarding Clearml HyperParameterOptimizer
I have a model that has as input Model(input = list, num_features = len(list)) , I want to do a hyperparameter search over my input. I have a list of lists containing the parameters I want it to try. How do I dynamically change my num_features parameter in the hyperparameteroptimizer.

Currently i have something that looks like this:

feature_list = [[10,20], [30,40,50]] # These are the two list I want to search through
num_features_list = [2,3] # The associated len of the list

optimizer = HyperparameterOptimizer(
hyper_parameters = [
DiscreteParameterRange('input', feature_list)
DiscreteParameterRange('num_features', num_features_list)
]

But this wont work because it will create an experiment with list [10,20] and num_features = 3 which is not what I want. Thank you very much

  
  
Posted 3 months ago
Votes Newest

Answers


Hi @<1547028031053238272:profile|MassiveGoldfish6> ! You can't enforce relationships between parameters. There are 2 things you could do:

  • prune tasks with bad relationships
  • create only one DiscreteParamterRange which is a list oh all the valid combinations of feature_list and num_features , but this would also require the task to optimize to support it
  
  
Posted 3 months ago
236 Views
1 Answer
3 months ago
3 months ago
Tags
Similar posts