def test_train_one_head(self):
head1 = multi_label_head.MultiLabelHead(n_classes=2, name='head1')
multi_head = multi_head_lib.MultiHead([head1])
logits = {
'head1': np.array([[-10., 10.], [-15., 10.]], dtype=np.float32)
}
expected_probabilities = {
'head1': nn.sigmoid(logits['head1']),
}
labels = {'head1': np.array([[1, 0], [1, 1]], dtype=np.int64)}
features = {'x': np.array(((42, ), ), dtype=np.int32)}
# For large logits, sigmoid cross entropy loss is approximated as:
# loss = labels * (logits < 0) * (-logits) +
# (1 - labels) * (logits > 0) * logits =>
# expected_unweighted_loss = [[10., 10.], [15., 0.]]
# loss = ((10 + 10) / 2 + (15 + 0) / 2) / 2 = 8.75
expected_loss = 8.75
tol = 1e-3
loss = multi_head.loss(logits=logits,
labels=labels,
features=features,
mode=model_fn.ModeKeys.TRAIN)
self.assertAllClose(expected_loss,
self.evaluate(loss),
rtol=tol,
atol=tol)
if context.executing_eagerly():
return
expected_train_result = 'my_train_op'
def _train_op_fn(loss):
return string_ops.string_join([
constant_op.constant(expected_train_result),
string_ops.as_string(loss, precision=3)
])
spec = multi_head.create_estimator_spec(features=features,
mode=model_fn.ModeKeys.TRAIN,
logits=logits,
labels=labels,
train_op_fn=_train_op_fn)
self.assertIsNotNone(spec.loss)
self.assertEqual({}, spec.eval_metric_ops)
self.assertIsNotNone(spec.train_op)
self.assertIsNone(spec.export_outputs)
test_lib._assert_no_hooks(self, spec)
# Assert predictions, loss, train_op, and summaries.
with self.cached_session() as sess:
test_lib._initialize_variables(self, spec.scaffold)
self.assertIsNotNone(spec.scaffold.summary_op)
loss, train_result, summary_str, predictions = sess.run(
(spec.loss, spec.train_op, spec.scaffold.summary_op,
spec.predictions))
self.assertAllClose(
logits['head1'],
predictions[('head1', prediction_keys.PredictionKeys.LOGITS)])
self.assertAllClose(
expected_probabilities['head1'],
predictions[('head1',
prediction_keys.PredictionKeys.PROBABILITIES)])
self.assertAllClose(expected_loss, loss, rtol=tol, atol=tol)
self.assertEqual(
six.b('{0:s}{1:.3f}'.format(expected_train_result,
expected_loss)), train_result)
test_lib._assert_simple_summaries(
self, {
metric_keys.MetricKeys.LOSS: expected_loss,
metric_keys.MetricKeys.LOSS + '/head1': expected_loss,
}, summary_str, tol)
def test_train_with_regularization_losses(self):
head1 = multi_label_head.MultiLabelHead(n_classes=2, name='head1')
head2 = multi_label_head.MultiLabelHead(n_classes=3, name='head2')
multi_head = multi_head_lib.MultiHead([head1, head2],
head_weights=[1., 2.])
logits = {
'head1':
np.array([[-10., 10.], [-15., 10.]], dtype=np.float32),
'head2':
np.array([[20., -20., 20.], [-30., 20., -20.]], dtype=np.float32),
}
expected_probabilities = {
'head1': nn.sigmoid(logits['head1']),
'head2': nn.sigmoid(logits['head2']),
}
labels = {
'head1': np.array([[1, 0], [1, 1]], dtype=np.int64),
'head2': np.array([[0, 1, 0], [1, 1, 0]], dtype=np.int64),
}
features = {'x': np.array(((42, ), ), dtype=np.int32)}
regularization_losses = [1.5, 0.5]
# For large logits, sigmoid cross entropy loss is approximated as:
# loss = labels * (logits < 0) * (-logits) +
# (1 - labels) * (logits > 0) * logits =>
# head1: expected_unweighted_loss = [[10., 10.], [15., 0.]]
# loss1 = ((10 + 10) / 2 + (15 + 0) / 2) / 2 = 8.75
# head2: expected_unweighted_loss = [[20., 20., 20.], [30., 0., 0]]
# loss2 = ((20 + 20 + 20) / 3 + (30 + 0 + 0) / 3) / 2 = 15
# Average over classes, weighted sum over batch and heads.
# weights = [1., 2.]
# merged_training_loss = 1. * loss1 + 2. * loss2
# training_loss = merged_training_loss + regularization_loss
# = 1. * loss1 + 2. * loss2 + sum([1.5, 0.5])
expected_loss_head1 = 8.75
expected_loss_head2 = 15.0
expected_regularization_loss = 2.
# training loss.
expected_loss = (1. * expected_loss_head1 + 2. * expected_loss_head2 +
expected_regularization_loss)
tol = 1e-3
loss = multi_head.loss(logits=logits,
labels=labels,
features=features,
mode=model_fn.ModeKeys.TRAIN,
regularization_losses=regularization_losses)
self.assertAllClose(expected_loss,
self.evaluate(loss),
rtol=tol,
atol=tol)
if context.executing_eagerly():
return
keys = metric_keys.MetricKeys
expected_train_result = 'my_train_op'
def _train_op_fn(loss):
return string_ops.string_join([
constant_op.constant(expected_train_result),
string_ops.as_string(loss, precision=3)
])
spec = multi_head.create_estimator_spec(
features=features,
mode=model_fn.ModeKeys.TRAIN,
logits=logits,
labels=labels,
train_op_fn=_train_op_fn,
regularization_losses=regularization_losses)
self.assertIsNotNone(spec.loss)
self.assertEqual({}, spec.eval_metric_ops)
self.assertIsNotNone(spec.train_op)
self.assertIsNone(spec.export_outputs)
test_lib._assert_no_hooks(self, spec)
# Assert predictions, loss, train_op, and summaries.
with self.cached_session() as sess:
test_lib._initialize_variables(self, spec.scaffold)
self.assertIsNotNone(spec.scaffold.summary_op)
loss, train_result, summary_str, predictions = sess.run(
(spec.loss, spec.train_op, spec.scaffold.summary_op,
spec.predictions))
self.assertAllClose(
logits['head1'],
predictions[('head1', prediction_keys.PredictionKeys.LOGITS)])
self.assertAllClose(
expected_probabilities['head1'],
predictions[('head1',
prediction_keys.PredictionKeys.PROBABILITIES)])
self.assertAllClose(
logits['head2'],
predictions[('head2', prediction_keys.PredictionKeys.LOGITS)])
self.assertAllClose(
expected_probabilities['head2'],
predictions[('head2',
prediction_keys.PredictionKeys.PROBABILITIES)])
self.assertAllClose(expected_loss, loss, rtol=tol, atol=tol)
self.assertEqual(
six.b('{0:s}{1:.3f}'.format(expected_train_result,
expected_loss)), train_result)
test_lib._assert_simple_summaries(
self, {
keys.LOSS_REGULARIZATION: expected_regularization_loss,
keys.LOSS: expected_loss,
keys.LOSS + '/head1': expected_loss_head1,
keys.LOSS + '/head2': expected_loss_head2,
}, summary_str, tol)
def test_eval_two_heads_with_weights(self):
head1 = multi_label_head.MultiLabelHead(n_classes=2, name='head1')
head2 = multi_label_head.MultiLabelHead(n_classes=3, name='head2')
multi_head = multi_head_lib.MultiHead([head1, head2],
head_weights=[1., 2.])
logits = {
'head1':
np.array([[-10., 10.], [-15., 10.]], dtype=np.float32),
'head2':
np.array([[20., -20., 20.], [-30., 20., -20.]], dtype=np.float32),
}
labels = {
'head1': np.array([[1, 0], [1, 1]], dtype=np.int64),
'head2': np.array([[0, 1, 0], [1, 1, 0]], dtype=np.int64),
}
features = {'x': np.array(((42, ), ), dtype=np.int32)}
# For large logits, sigmoid cross entropy loss is approximated as:
# loss = labels * (logits < 0) * (-logits) +
# (1 - labels) * (logits > 0) * logits =>
# head1: expected_unweighted_loss = [[10., 10.], [15., 0.]]
# loss = ((10 + 10) / 2 + (15 + 0) / 2) / 2 = 8.75
# head2: expected_unweighted_loss = [[20., 20., 20.], [30., 0., 0]]
# loss = ((20 + 20 + 20) / 3 + (30 + 0 + 0) / 3) / 2 = 15
expected_loss_head1 = 8.75
expected_loss_head2 = 15.
expected_loss = 1. * expected_loss_head1 + 2. * expected_loss_head2
tol = 1e-3
keys = metric_keys.MetricKeys
expected_metrics = {
keys.LOSS + '/head1': expected_loss_head1,
keys.LOSS + '/head2': expected_loss_head2,
# Average loss over examples.
keys.LOSS_MEAN + '/head1': expected_loss_head1,
keys.LOSS_MEAN + '/head2': expected_loss_head2,
# auc and auc_pr cannot be reliably calculated for only 4-6 samples, but
# this assert tests that the algorithm remains consistent.
# TODO(yhliang): update metrics
# keys.AUC + '/head1': 0.1667,
# keys.AUC + '/head2': 0.3333,
# keys.AUC_PR + '/head1': 0.6667,
# keys.AUC_PR + '/head2': 0.5000,
}
if context.executing_eagerly():
loss = multi_head.loss(logits,
labels,
features=features,
mode=model_fn.ModeKeys.EVAL)
self.assertIsNotNone(loss)
self.assertAllClose(expected_loss, loss, rtol=tol, atol=tol)
eval_metrics = multi_head.metrics()
updated_metrics = multi_head.update_metrics(
eval_metrics, features, logits, labels)
self.assertItemsEqual(expected_metrics.keys(),
updated_metrics.keys())
self.assertAllClose(
expected_metrics,
{k: updated_metrics[k].result()
for k in updated_metrics},
rtol=tol,
atol=tol)
return
spec = multi_head.create_estimator_spec(features=features,
mode=model_fn.ModeKeys.EVAL,
logits=logits,
labels=labels)
# Assert spec contains expected tensors.
self.assertIsNotNone(spec.loss)
self.assertItemsEqual(expected_metrics.keys(),
spec.eval_metric_ops.keys())
self.assertIsNone(spec.train_op)
self.assertIsNone(spec.export_outputs)
test_lib._assert_no_hooks(self, spec)
# Assert predictions, loss, and metrics.
with self.cached_session() as sess:
test_lib._initialize_variables(self, spec.scaffold)
self.assertIsNone(spec.scaffold.summary_op)
value_ops = {
k: spec.eval_metric_ops[k][0]
for k in spec.eval_metric_ops
}
update_ops = {
k: spec.eval_metric_ops[k][1]
for k in spec.eval_metric_ops
}
loss, _ = sess.run((spec.loss, update_ops))
self.assertAllClose(expected_loss, loss, rtol=tol, atol=tol)
# Check results of value ops (in `metrics`).
self.assertAllClose(expected_metrics,
{k: value_ops[k].eval()
for k in value_ops},
rtol=tol,
atol=tol)
