hiro/ray
1
0
Fork 0
mirror of https://github.com/vale981/ray synced 2025-03-06 10:31:39 -05:00
Code Issues Projects Releases Packages Wiki Activity Actions

[minor][tune] update pbt docs (#5420)

Browse source
This commit is contained in:
Richard Liaw 2019-08-11 12:39:54 -07:00 • committed by GitHub
parent b1e010feec
commit cff72d1a54
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 3 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
python/ray/tune/ray_trial_executor.py
View file

@ -97,7 +97,7 @@ class RayTrialExecutor(TrialExecutor):
trial.runner = existing_runner trial.runner = existing_runner
if not self.reset_trial(trial, trial.config, trial.experiment_tag): if not self.reset_trial(trial, trial.config, trial.experiment_tag):
raise AbortTrialExecution( raise AbortTrialExecution(
"Trial runner reuse requires reset_trial() to be " "Trainable runner reuse requires reset_config() to be "
"implemented and return True.") "implemented and return True.")
return existing_runner return existing_runner

4
python/ray/tune/schedulers/pbt.py
View file

@ -109,7 +109,7 @@ class PopulationBasedTraining(FIFOScheduler):
This Tune PBT implementation considers all trials added as part of the This Tune PBT implementation considers all trials added as part of the
PBT population. If the number of trials exceeds the cluster capacity, PBT population. If the number of trials exceeds the cluster capacity,
they will be time-multiplexed as to balance training progress across the they will be time-multiplexed as to balance training progress across the
population. population. To run multiple trials, use `tune.run(num_samples=<int>)`.
Args: Args:
time_attr (str): The training result attr to use for comparing time. time_attr (str): The training result attr to use for comparing time.
@ -162,7 +162,7 @@ class PopulationBasedTraining(FIFOScheduler):
>>> # 10 -> 1 or 10 -> 100. Resampling will choose at random. >>> # 10 -> 1 or 10 -> 100. Resampling will choose at random.
>>> "factor_2": [1, 10, 100, 1000, 10000], >>> "factor_2": [1, 10, 100, 1000, 10000],
>>> }) >>> })
>>> run_experiments({...}, scheduler=pbt) >>> tune.run({...}, num_samples=8, scheduler=pbt)
""" """
def __init__(self, def __init__(self,