def test_tf_superloss_hinge(self):
true = tf.constant([[-1, 1, 1, -1], [1, 1, 1, 1], [-1, -1, 1, -1], [1, -1, -1, -1]])
pred = tf.constant([[0.1, 0.9, 0.05, 0.05], [0.1, -0.2, 0.0, -0.7], [0.0, 0.15, 0.8, 0.05],
[1.0, -1.0, -1.0, -1.0]])
sl = SuperLoss(Hinge(inputs=('x1', 'x2'), outputs='x'))
sl.build('tf', None)
output = sl.forward(data=[pred, true], state=self.state)
self.assertTrue(np.allclose(output.numpy(), -0.072016776))
def test_torch_superloss_hinge(self):
true = torch.tensor([[-1, 1, 1, -1], [1, 1, 1, 1], [-1, -1, 1, -1],
[1, -1, -1, -1]]).to("cuda:0" if torch.cuda.is_available() else "cpu")
pred = torch.tensor([[0.1, 0.9, 0.05, 0.05], [0.1, -0.2, 0.0, -0.7], [0.0, 0.15, 0.8, 0.05],
[1.0, -1.0, -1.0, -1.0]]).to("cuda:0" if torch.cuda.is_available() else "cpu")
sl = SuperLoss(Hinge(inputs=('x1', 'x2'), outputs='x'))
sl.build('torch', "cuda:0" if torch.cuda.is_available() else "cpu")
output = sl.forward(data=[pred, true], state=self.state)
self.assertTrue(np.allclose(output.to("cpu").numpy(), -0.072016776))
def test_torch_superloss_binary_ce(self):
sl = SuperLoss(CrossEntropy(inputs=['y_pred', 'y'], outputs='ce'))
sl.build(framework="torch",
device="cuda:0" if torch.cuda.is_available() else "cpu")
output = sl.forward(
data=[self.torch_pred_binary, self.torch_true_binary],
state=self.state)
self.assertTrue(
np.allclose(output.detach().to("cpu").numpy(), -0.0026238672))
def test_tf_superloss_sparse_categorical_ce(self):
sl = SuperLoss(CrossEntropy(inputs=['y_pred', 'y'], outputs='ce'))
sl.build(framework="tf", device=None)
output = sl.forward(data=[self.tf_pred_sparse, self.tf_true_sparse], state=self.state)
self.assertTrue(np.allclose(output.numpy(), -0.024740249))