hyperparam_tuning.py

input_data = {"_train_sample":'/dbfs/mnt/blogs_pl/taxi_fare_feature_eng_train_sample6', "_validate_sample":'/dbfs/mnt/blogs_pl/taxi_fare_feature_eng_validate_sample6',"_test_sample":'/dbfs/mnt/blogs_pl/taxi_fare_feature_eng_test_sample6'}
activation_function = ['relu', 'tanh', 'sigmoid']

for n in range (3):
  learning_rate = n+1/1000
  batch_size = 512 * (n+1)
  for act in activation_function:
    input_params = { "_learning_rate":learning_rate, "_steps":100000, "_batch_size":batch_size, "_dataset_size":4000000,\
                    "_model_dir":'/dbfs/tmp/models', "_activation_function":act,\
                    "_checkpoints_steps":5000,"_output_path":'/dbfs/mnt/blogs_pl/output1'}

    nyt= NYorkTaxiFairPrediction.new_instance(input_params,input_data)
    (experimentID, runID) =  nyt.mlflow_run(NYorkTaxiFairPrediction.random_key(10))
    print("MLflow Run for NYorkTaxiFairPrediction completed with run_id {} and experiment_id {}".format(runID, experimentID))
    print("-" * 100)