def generate_parameters(self, parameter_id, **kwargs):
"""
This function will returns a dict of trial (hyper-)parameters.
If no trial configration for now, self.credit plus 1 to send the config later
Parameters
----------
parameter_id : int
Returns
-------
dict
One newly generated configuration.
"""
if not self.population:
raise RuntimeError('The population is empty')
if self.num_running_trials >= self.population_size:
logger.warning(
"No enough trial config, population_size is suggested to be larger than trialConcurrency"
)
self.credit += 1
self.param_ids.append(parameter_id)
raise nni.NoMoreTrialError('no more parameters now.')
return self._generate_individual(parameter_id)
def generate_parameters(self, parameter_id, **kwargs):
"""
Generate parameters, if no trial configration for now, self.credit plus 1 to send the config later
Parameters
----------
parameter_id : int
Unique identifier for requested hyper-parameters.
This will later be used in :meth:`receive_trial_result`.
**kwargs
Not used
Returns
-------
dict
One newly generated configuration
"""
if self.pos == self.population_size - 1:
logger.debug('Credit added by one in parameters request')
self.credit += 1
self.param_ids.append(parameter_id)
raise nni.NoMoreTrialError('No more parameters now.')
self.pos += 1
trial_info = self.population[self.pos]
trial_info.parameter_id = parameter_id
self.running[parameter_id] = trial_info
logger.info('Generate parameter : %s', trial_info.hyper_parameters)
return trial_info.hyper_parameters
def generate_parameters(self, parameter_id, **kwargs):
"""
Generate parameters for one trial.
Parameters
----------
parameter_id : int
The id for the generated hyperparameter
**kwargs
Not used
Returns
-------
dict
One configuration from the expanded search space.
Raises
------
NoMoreTrialError
If all the configurations has been sent, raise :class:`~nni.NoMoreTrialError`.
"""
self.count += 1
while self.count <= len(self.expanded_search_space) - 1:
_params_tuple = convert_dict2tuple(
copy.deepcopy(self.expanded_search_space[self.count]))
if _params_tuple in self.supplement_data:
self.count += 1
else:
return self.expanded_search_space[self.count]
raise nni.NoMoreTrialError('no more parameters now.')
def generate_parameters(self, parameter_id, **kwargs):
"""
Generate parameters, if no trial configration for now, self.credit plus 1 to send the config later
Parameters
----------
parameter_id : int
Unique identifier for requested hyper-parameters.
This will later be used in :meth:`receive_trial_result`.
**kwargs
Not used
Returns
-------
dict
One newly generated configuration
"""
if self.first_inf:
self.trials_result = [None for _ in range(self.inf_batch_size)]
mb_obs, mb_actions, mb_values, mb_neglogpacs, mb_dones, last_values = self.model.inference(self.inf_batch_size)
self.trials_info = TrialsInfo(mb_obs, mb_actions, mb_values, mb_neglogpacs,
mb_dones, last_values, self.inf_batch_size)
self.first_inf = False
trial_info_idx, actions = self.trials_info.get_next()
if trial_info_idx is None:
logger.debug('Credit added by one in parameters request')
self.credit += 1
self.param_ids.append(parameter_id)
raise nni.NoMoreTrialError('no more parameters now.')
self.running_trials[parameter_id] = trial_info_idx
new_config = self._actions_to_config(actions)
return new_config
def generate_parameters(self, parameter_id):
self.count += 1
while (self.count <= len(self.expanded_search_space)-1):
_params_tuple = convert_dict2tuple(self.expanded_search_space[self.count])
if _params_tuple in self.supplement_data:
self.count += 1
else:
return self.expanded_search_space[self.count]
raise nni.NoMoreTrialError('no more parameters now.')
def generate_parameters(self, parameter_id, **kwargs):
"""Method which provides one set of hyper-parameters.
Override of the abstract method in :class:`~nni.tuner.Tuner`.
"""
if self.serve_list:
self.wait_dict[parameter_id] = self.serve_list.pop()
return self.wait_dict[parameter_id].pick_out()
else:
self.request_list.append(parameter_id)
raise nni.NoMoreTrialError('no more parameters now.')
def generate_parameters(self, parameter_id):
"""Returns a dict of trial (hyper-)parameters, as a serializable object.
Parameters
----------
parameter_id : int
"""
self.count += 1
if self.count > len(self.values) - 1:
raise nni.NoMoreTrialError('no more parameters now.')
return self.values[self.count]
def generate_parameters(self, *args, **kwargs):
while True:
params = self._suggest()
if params is None:
raise nni.NoMoreTrialError('Search space fully explored')
params = deformat_parameters(params, self.space)
params_str = nni.dump(params, sort_keys=True)
if params_str not in self.history:
self.history.add(params_str)
return params
def generate_parameters(self, parameter_id, **kwargs):
"""Returns a dict of trial (hyper-)parameters, as a serializable object.
Parameters
----------
parameter_id : int
Returns
-------
dict
A candidate parameter group.
"""
self._count += 1
if self._count > len(self._values) - 1:
raise nni.NoMoreTrialError('no more parameters now.')
return self._values[self._count]
def __call__(
self,
formatted_parameters: FormattedParameters) -> FormattedParameters:
if self._never_dup or self._not_dup(formatted_parameters):
return formatted_parameters
if self._grid_search is None:
_logger.info(
f'Tuning algorithm generated duplicate parameter: {formatted_parameters}'
)
_logger.info(f'Use grid search for deduplication.')
self._init_grid_search()
while True:
new = self._grid_search._suggest() # type: ignore
if new is None:
raise nni.NoMoreTrialError()
if self._not_dup(new):
return new
def generate_parameters(self, parameter_id, **kwargs):
"""
generate parameters, if no trial configration for now, self.credit plus 1 to send the config later
"""
if self.first_inf:
self.trials_result = [None for _ in range(self.inf_batch_size)]
mb_obs, mb_actions, mb_values, mb_neglogpacs, mb_dones, last_values = self.model.inference(self.inf_batch_size)
self.trials_info = TrialsInfo(mb_obs, mb_actions, mb_values, mb_neglogpacs,
mb_dones, last_values, self.inf_batch_size)
self.first_inf = False
trial_info_idx, actions = self.trials_info.get_next()
if trial_info_idx is None:
self.credit += 1
self.param_ids.append(parameter_id)
raise nni.NoMoreTrialError('no more parameters now.')
self.running_trials[parameter_id] = trial_info_idx
new_config = self._actions_to_config(actions)
return new_config
def generate_parameters(self, parameter_id, trial_job_id=None, pos=None):
if pos is None:
if not self.bucket:
if self.num_completed_jobs < self.total_steps:
raise nni.NoMoreTrialError('no more parameters now.')
else:
self.generate_one_epoch_parameters()
pos = self.bucket.pop()
logger.info('current bucket: %s', self.bucket)
logger.info('current pos: %s', pos)
self.parameter_id2pos[parameter_id] = pos
current_arc_code = self.child_arc[pos]
start_idx = 0
current_config = dict()
def onehot2list(l):
return [idx for idx, val in enumerate(l) if val == 1]
for layer_id, (layer_name, info) in enumerate(self.search_space):
mutable_block = info['mutable_block']
if mutable_block not in current_config:
current_config[mutable_block] = dict()
layer_choice_idx = current_arc_code[start_idx]
if layer_id != 0:
input_start = start_idx + 1
else:
input_start = start_idx
inputs_idxs = current_arc_code[input_start:input_start + layer_id]
inputs_idxs = onehot2list(inputs_idxs)
current_config[mutable_block][layer_name] = dict()
current_config[mutable_block][layer_name]['chosen_layer'] = info[
'layer_choice'][layer_choice_idx]
current_config[mutable_block][layer_name]['chosen_inputs'] = [
info['optional_inputs'][ipi] for ipi in inputs_idxs
]
start_idx += 1 + layer_id
return current_config
def generate_parameters(self, parameter_id, **kwargs):
"""
Generate one instance of hyperparameters (i.e., one configuration).
Get one from SMAC3's ``challengers``.
Parameters
----------
parameter_id : int
Unique identifier for requested hyper-parameters. This will later be used in :meth:`receive_trial_result`.
**kwargs
Not used
Returns
-------
dict
One newly generated configuration
"""
if self.first_one:
init_challenger = self.smbo_solver.nni_smac_start()
self.total_data[parameter_id] = init_challenger
return self.param_postprocess(init_challenger.get_dictionary())
else:
challengers = self.smbo_solver.nni_smac_request_challengers()
challengers_empty = True
for challenger in challengers:
challengers_empty = False
if self.dedup:
match = [v for k, v in self.total_data.items() \
if v.get_dictionary() == challenger.get_dictionary()]
if match:
continue
self.total_data[parameter_id] = challenger
return self.param_postprocess(challenger.get_dictionary())
assert challengers_empty is False, 'The case that challengers is empty is not handled.'
self.logger.info('In generate_parameters: No more new parameters.')
raise nni.NoMoreTrialError('No more new parameters.')
def generate_parameters(self, parameter_id):
self.count += 1
if self.count > len(self.expanded_search_space) - 1:
raise nni.NoMoreTrialError('no more parameters now.')
return self.expanded_search_space[self.count]
def generate_parameters(self, parameter_id):
self.count += 1
if self.count > len(self.values) - 1:
raise nni.NoMoreTrialError('no more parameters now.')
return self.values[self.count]
