Verbose setting for automl

[lesshaste]

Using the latest version 0.1.0 from pypi and the example from http://automl.github.io/auto-sklearn/stable/, there is no output at all while it is running apart from

You are already timing task: index_run4 You are already timing task: index_run5 You are already timing task: index_run8 You are already timing task: index_run8 You are already timing task: index_run8 You are already timing task: index_run8 You are already timing task: index_run8 You are already timing task: index_run8 You are already timing task: index_run8 You are already timing task: index_run8 You are already timing task: index_run8 You are already timing task: index_run8 You are already timing task: index_run8 You are already timing task: index_run8 You are already timing task: index_run8 You are already timing task: index_run8 You are already timing task: index_run8 You are already timing task: index_run8 You are already timing task: index_run8 You are already timing task: index_run8 n_components is too large: it will be set to 64 [...]

It would be great if there were a verbose option. The most obvious things one would like to see are:

a) What is in the current ensemble? b) What transformations have been applied to the original data? c) What is the current best score?

[zygmuntz]

As a work-around, you can see the log in /tmp, something like $ less +F /tmp/autosklearn_tmp_2876_7305/AutoML\(1\)\:2c1db42a90a9a269fcb61b36f65577c2.log.

[lesshaste]

Thanks. Just to complete my copy and paste above, the output currently terminates with:

You are already timing task: index_run8
/home/user/.local/lib/python2.7/site-packages/sklearn/svm/base.py:547: ChangedBehaviorWarning: The decision_function_shape default value will change from 'ovo' to 'ovr' in 0.18. This will change the shape of the decision function returned by SVC.
  "SVC.", ChangedBehaviorWarning)
/home/user/.local/lib/python2.7/site-packages/sklearn/discriminant_analysis.py:515: RuntimeWarning: overflow encountered in exp
  np.exp(prob, prob)
/home/user/.local/lib/python2.7/site-packages/sklearn/feature_selection/univariate_selection.py:523: RuntimeWarning: invalid value encountered in less
  return self.pvalues_ < self.alpha
/home/user/.local/lib/python2.7/site-packages/sklearn/feature_selection/univariate_selection.py:578: RuntimeWarning: invalid value encountered in less_equal
  * np.arange(n_features)]
/home/user/.local/lib/python2.7/site-packages/sklearn/feature_selection/univariate_selection.py:581: RuntimeWarning: invalid value encountered in less_equal
  return self.pvalues_ <= selected.max()
('Accuracy score', 0.99111111111111116)

[mfeurer]

These are warnings during the call to predict. Actually it would make sense to wrap them, too. Thanks for mentioning this.