def get_query_with_settings(settings):
"""Function to get the query method"""
method = settings.query_strategy
if method in ['rand_max', 'rand_max_sampling']:
settings.query_kwargs['rand_max_frac'] = 0.05
settings.query_kwargs = _unsafe_dict_update(
settings.query_kwargs, settings.query_param)
return get_query_strategy(method)
def lstm_base_model_defaults(settings, verbose=1):
""" Set the lstm model defaults. """
model_kwargs = {}
model_kwargs['backwards'] = True
model_kwargs['dropout'] = 0.4
model_kwargs['optimizer'] = "rmsprop"
model_kwargs['max_sequence_length'] = 1000
model_kwargs['verbose'] = verbose
model_kwargs['lstm_out_width'] = 20
model_kwargs['dense_width'] = 128
upd_param = _unsafe_dict_update(model_kwargs, settings.model_param)
settings.model_param = upd_param
return upd_param
def lstm_fit_defaults(settings, verbose=1):
""" Set the fit defaults and merge them with custom settings. """
# arguments to pass to the fit
fit_kwargs = {}
fit_kwargs['batch_size'] = 32
fit_kwargs['epochs'] = 10
fit_kwargs['verbose'] = verbose
fit_kwargs['shuffle'] = False
fit_kwargs['class_weight_inc'] = 30.0
settings.fit_kwargs = _unsafe_dict_update(fit_kwargs, settings.fit_param)
_set_class_weight(fit_kwargs.pop('class_weight_inc'), fit_kwargs)
return settings.fit_kwargs
def get_query_strategy(settings):
"""Function to get the query method"""
method = settings.query_strategy
if method in ['max', 'max_sampling']:
return max_sampling, "Maximum inclusion sampling"
if method in ['rand_max', 'rand_max_sampling']:
settings.query_kwargs['rand_max_frac'] = 0.05
settings.query_kwargs = _unsafe_dict_update(settings.query_kwargs,
settings.query_param)
return rand_max_sampling, "Mix of random and max inclusion sampling"
elif method in ['lc', 'sm', 'uncertainty', 'uncertainty_sampling']:
return uncertainty_sampling, 'Least confidence / Uncertainty sampling'
elif method == 'random':
return random_sampling, 'Random'
else:
raise ValueError(f"Query strategy '{method}' not found.")
def __init__(self, balance_kwargs):
self.balance_kwargs = self.default_kwargs()
self.balance_kwargs = _unsafe_dict_update(self.balance_kwargs,
balance_kwargs)
def __init__(self, param):
self.name = "base"
self.param = _unsafe_dict_update(self.default_param(), param)
