def prepare(self, params, train_size):
self.params = params
self.train_size = train_size
self.resampler = WeightedResampler(train_size)
self.D = self.resampler.weights
self.alphas = []
self.member_number = 0
def prepare(self, params, dataset):
self.params = params
self.dataset = dataset
self.resampler = WeightedResampler(dataset)
self.D = self.resampler.weights
self.alphas = []
self.member_number = 0
def prepare(self, params, dataset):
self.params = params
self.dataset = dataset
self.resampler = WeightedResampler(dataset)
self.D = self.resampler.weights # <--- sets the initial weights (step 1 in [1])
self.alphas = []
self.member_number = 0
def prepare(self, params, dataset):
self.params = params
self.dataset = dataset
self.resampler = WeightedResampler(dataset)
self.D = self.resampler.weights
self.weights = None
self.member_number = 0
self.alphas = []
model_yaml = read_yaml_file(params.model_file)
self.model_config = keras.models.model_from_yaml(
model_yaml).get_config()
def test_weighted_resampler(self):
r = WeightedResampler(self.dataset)
print numpy.asarray(r.make_new_train(10))
r.update_weights(self.weights)
print numpy.asarray(r.make_new_train(10))