def test_scale_std_of_dataset(self): ds = { 'training': (self.X, FramewiseTargets(self.X)), 'validation': (self.X + 1, FramewiseTargets(self.X)), 'test': (self.X * 2, FramewiseTargets(self.X)) } assert_allclose(get_stds(ds['training'][0]), [6.90410506, 6.90410506]) assert_allclose(get_stds(ds['validation'][0]), [6.90410506, 6.90410506]) assert_allclose(get_stds(ds['test'][0]), [13.80821012, 13.80821012]) scale_std_of_dataset(ds) assert_allclose(get_stds(ds['training'][0]), [1., 1.], atol=1e-6) assert_allclose(get_stds(ds['validation'][0]), [1., 1.], atol=1e-6) assert_allclose(get_stds(ds['test'][0]), [2., 2.], atol=1e-6)
def test_scale_std_of_dataset_masked(self): ds = { 'training': (self.X, FramewiseTargets(self.X, self.M)), 'validation': (self.X + 1, FramewiseTargets(self.X, self.M)), 'test': (self.X * 2, FramewiseTargets(self.X, self.M)) } assert_allclose(get_stds(ds['training'][0], self.M), [6.27992834, 6.27992834]) assert_allclose(get_stds(ds['validation'][0], self.M), [6.27992834, 6.27992834]) assert_allclose(get_stds(ds['test'][0], self.M), [12.55985668, 12.55985668]) scale_std_of_dataset(ds) assert_allclose(get_stds(ds['training'][0], self.M), [1., 1.]) assert_allclose(get_stds(ds['validation'][0], self.M), [1., 1.]) assert_allclose(get_stds(ds['test'][0], self.M), [2., 2.])