def test_unequal_weights_reduction_2d(self):
"""Testing unequal weights reductions for 2D data"""
y_true = [[1, 2], [0, 2]]
y_pred = [[[0.05, 0.95, 0], [0.1, 0.8, 0.1]],
[[0.1, 0.2, 0.7], [0.3, 0.5, 0.2]]]
sample_weight = np.take(self.class_weights, y_true)
scce = self.scce(y_true, y_pred, sample_weight=sample_weight).numpy()
self.sfl_unequal = BalancedSparseFocalLoss(
self.class_weights,
gamma=self.gamma,
reduction=losses.Reduction.SUM)
scce1 = scce * \
(1 - np.where(get_one_hot(y_true), y_pred, 0).sum(axis=-1))\
**self.gamma
scce1 = scce1.sum()
sfl1 = self.sfl_unequal(y_true, y_pred).numpy()
np.testing.assert_allclose(scce1, sfl1, rtol=1e-6)
self.sfl_unequal = BalancedSparseFocalLoss(
self.class_weights,
gamma=self.gamma,
reduction=losses.Reduction.NONE)
scce2 = scce * \
(1 - np.where(get_one_hot(y_true), y_pred, 0).sum(axis=-1))\
**self.gamma
sfl2 = self.sfl_unequal(y_true, y_pred).numpy()
np.testing.assert_allclose(scce2, sfl2, rtol=1e-5)
def setUp(self):
"""Setup shared by all tests"""
self.scce = losses.SparseCategoricalCrossentropy(
reduction=losses.Reduction.NONE)
self.gamma = 4.0
self.sfl_equal = BalancedSparseFocalLoss([1, 1, 1], gamma=self.gamma)
self.class_weights = [0.2, 0.3, 0.5]
self.sfl_unequal = BalancedSparseFocalLoss(self.class_weights,
gamma=self.gamma)
def test_equal_weights_reduction_1d(self):
"""Testing equal weights reductions for 1D data"""
y_true = [1, 2]
y_pred = [[0.05, 0.95, 0], [0.1, 0.8, 0.1]]
scce = self.scce(y_true, y_pred).numpy()
self.sfl_equal = BalancedSparseFocalLoss(
[1, 1, 1], gamma=self.gamma, reduction=losses.Reduction.SUM)
scce1 = scce * \
(1 - np.where(get_one_hot(y_true), y_pred, 0).sum(axis=-1))\
**self.gamma
scce1 = scce1.sum()
sfl1 = self.sfl_equal(y_true, y_pred).numpy()
np.testing.assert_allclose(scce1, sfl1, rtol=1e-6)
self.sfl_equal = BalancedSparseFocalLoss(
[1, 1, 1], gamma=self.gamma, reduction=losses.Reduction.NONE)
scce2 = scce * \
(1 - np.where(get_one_hot(y_true), y_pred, 0).sum(axis=-1))\
**self.gamma
sfl2 = self.sfl_equal(y_true, y_pred).numpy()
np.testing.assert_allclose(scce2, sfl2, rtol=1e-5)
def test_equal_weights_logits_1d(self):
"""Testing equal weights logits for 1D data"""
y_true = [1, 2]
y_pred = [[-0.05, 0.3, 0.19], [0.2, -0.4, 0.12]]
self.scce = losses.SparseCategoricalCrossentropy(
from_logits=True, reduction=losses.Reduction.NONE)
scce = self.scce(y_true, y_pred).numpy()
self.sfl_equal = BalancedSparseFocalLoss([1, 1, 1],
gamma=self.gamma,
from_logits=True)
scce = scce * \
(1 - np.where(get_one_hot(y_true), softmax(y_pred), 0).sum(axis=-1))\
**self.gamma
scce = scce.mean()
sfl = self.sfl_equal(y_true, y_pred).numpy()
np.testing.assert_allclose(scce, sfl, rtol=1e-6)
def test_unequal_weights_logits_2d(self):
"""Testing unequal weights logits for 2D data"""
y_true = [[1, 2], [0, 2]]
y_pred = [[[-0.05, 0.3, 0.19], [0.2, -0.4, 0.12]],
[[-0.1, 0.22, -0.73], [0.23, -0.52, 0.2]]]
sample_weight = np.take(self.class_weights, y_true)
self.scce = losses.SparseCategoricalCrossentropy(
from_logits=True, reduction=losses.Reduction.NONE)
scce = self.scce(y_true, y_pred, sample_weight=sample_weight).numpy()
self.sfl_unequal = BalancedSparseFocalLoss(self.class_weights,
gamma=self.gamma,
from_logits=True)
scce = scce * \
(1 - np.where(get_one_hot(y_true), softmax(y_pred), 0).sum(axis=-1))\
**self.gamma
scce = scce.mean()
sfl = self.sfl_unequal(y_true, y_pred).numpy()
np.testing.assert_allclose(scce, sfl, rtol=1e-6)