Hyperparameter sweep YAML with configurable search space ranges

[CynthiaLuijkx]

Describe your suggestion

I'd like an example of how to do a sweep job while specifying the search space range from the inputs parameters. The example currently available currently does not cover this, and I've been having problems getting it to work myself.

I'd like to see something like this:

$schema: https://azuremlschemas.azureedge.net/latest/pipelineJob.schema.json
type: pipeline
display_name: pipeline_with_hyperparameter_sweep
description: Tune hyperparameters using TF component
settings:
    default_compute: azureml:cpu-cluster

inputs:
  c_value_min: 0.5
  c_value_max: 0.9

jobs:
  sweep_step:
    type: sweep
    inputs:
      data: 
        type: uri_file
        path: wasbs://[email protected]/iris.csv
      degree: 3
      gamma: "scale"
      shrinking: False
    outputs:
      model_output:
      test_data:
    sampling_algorithm: random
    trial: ./train.yml
    search_space:
      c_value:
        type: uniform
        min_value: ${{ parent.inputs.c_value_min }}
        max_value: ${{ parent.inputs.c_value_max }}
    objective:
      goal: minimize
      primary_metric: training_f1_score
    limits:
      max_total_trials: 5
      max_concurrent_trials: 3
      timeout: 7200

But then in a way that works; the method above unfortunately does not. The pipeline is submitted correctly, but in the sweep job itself the inputs provided are replaced with '0's as if they are missing. I have been unable to figure out the correct syntax to use here, or to find proper examples.

Additional details

No response