make default verbose level > 0 when using ray (#272)

* make default verbose level > 0 when using ray

* default hpo method when using ray

* bug fix: == -> =
This commit is contained in:
Chi Wang 2021-11-04 22:06:19 -07:00 committed by GitHub
parent 549a0dfb53
commit fc32eca24b
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23

View File

@ -1578,7 +1578,6 @@ class AutoML:
)
logger.info("List of ML learners in AutoML Run: {}".format(estimator_list))
self.estimator_list = estimator_list
self._hpo_method = hpo_method or ("cfo" if n_concurrent_trials == 1 else "bs")
self._state.time_budget = time_budget or 1e10
self._active_estimators = estimator_list.copy()
self._ensemble = ensemble
@ -1592,7 +1591,8 @@ class AutoML:
self._state.n_jobs = n_jobs
self._n_concurrent_trials = n_concurrent_trials
self._early_stop = early_stop
self._use_ray = use_ray or self._n_concurrent_trials > 1
self._use_ray = use_ray or n_concurrent_trials > 1
self._hpo_method = hpo_method or ("bs" if self._use_ray else "cfo")
if log_file_name:
with training_log_writer(log_file_name, append_log) as save_helper:
self._training_log = save_helper
@ -1709,7 +1709,7 @@ class AutoML:
resources_per_trial=resources_per_trial,
time_budget_s=self._state.time_budget,
num_samples=self._max_iter,
verbose=max(self.verbose - 3, 0),
verbose=max(self.verbose - 2, 0),
raise_on_failed_trial=False,
)
# logger.info([trial.last_result for trial in analysis.trials])
@ -2297,7 +2297,7 @@ class AutoML:
speed = delta_loss / delta_time
if speed:
estimated_cost = max(2 * gap / speed, estimated_cost)
estimated_cost == estimated_cost or 1e-9
estimated_cost = estimated_cost or 1e-9
inv.append(1 / estimated_cost)
else:
estimated_cost = self._eci[i]