def _batch_trials(batch_run):
proto_flag_vals = batch_run.batch_proto.get("flags")
batch_run = batch_util.batch_run()
max_trials = batch_run.get("max_trials") or DEFAULT_MAX_TRIALS
random_seed = batch_run.get("random_seed")
try:
return skopt_util.random_trials_for_flags(proto_flag_vals, max_trials,
random_seed)
except skopt_util.MissingSearchDimension as e:
skopt_util.missing_search_dim_error(proto_flag_vals)
except skopt_util.InvalidSearchDimension as e:
_search_dim_error(e)
def main():
batch_util.init_logging()
batch = batch_util.batch_run()
batch_flags = batch.get("flags")
max_trials = batch.get("max_trials") or DEFAULT_MAX_TRIALS
random_seed = batch.get("random_seed")
trials = hyperopt.Trials()
try:
hyperopt.fmin(
fn=_objective_fn(batch),
space=_space_for_flags(batch),
algo=_tpe_suggest,
max_evals=max_trials,
show_progressbar=False,
rstate=np.random.RandomState(random_seed),
trials=trials,
)
except hyperopt.exceptions.AllTrialsFailed:
pass
else:
_label_best(trials.best_trial)
def main():
batch_util.init_logging()
batch_run = batch_util.batch_run()
skopt_util.handle_seq_trials(batch_run, _suggest_x)
def main():
batch_util.init_logging()
batch_run = batch_util.batch_run()
trials = _batch_trials(batch_run)
batch_util.handle_trials(batch_run, trials)
from guild import batch_util
try:
batch_run = batch_util.batch_run()
except batch_util.CurrentRunNotBatchError:
print("This script must be run as a Guild optimizer")
else:
proto_flags = batch_run.batch_proto.get("flags", {})
print("Tune using proto flags: %s" % sorted(proto_flags.items()))
