Add more merging layers (#71)
* add base merge layer * format docstrings * add layer * add test cases for layer * Add import for layer * fix build function * add dynamic and static tests * fix pytest import * fix pytest decorator * remove batch size from dynamic shape test * fix keras reference * refactor test class * fix tf tests, and linting issues * add subtract layer * add tests for subtract layer * fix linting issues * add average layer * add maximum layer * dd minimum layer * add multiply layer * add tests for average, minimum, maximum, and multiply layers
This commit is contained in:
Aakash Kumar Nain
Francois Chollet
parent
bb6dc0fbf4
commit
50e106b56c
@ -6,6 +6,14 @@ from keras_core.layers.core.input_layer import InputLayer
|
||||
from keras_core.layers.layer import Layer
|
||||
from keras_core.layers.merging.add import Add
|
||||
from keras_core.layers.merging.add import add
|
||||
from keras_core.layers.merging.average import Average
|
||||
from keras_core.layers.merging.average import average
|
||||
from keras_core.layers.merging.maximum import Maximum
|
||||
from keras_core.layers.merging.maximum import maximum
|
||||
from keras_core.layers.merging.minimum import Minimum
|
||||
from keras_core.layers.merging.minimum import minimum
|
||||
from keras_core.layers.merging.multiply import Multiply
|
||||
from keras_core.layers.merging.multiply import multiply
|
||||
from keras_core.layers.merging.subtract import Subtract
|
||||
from keras_core.layers.merging.subtract import subtract
|
||||
from keras_core.layers.pooling.average_pooling1d import AveragePooling1D
|
||||
|
||||
69
keras_core/layers/merging/average.py
Normal file
69
keras_core/layers/merging/average.py
Normal file
@ -0,0 +1,69 @@
|
||||
from keras_core.api_export import keras_core_export
|
||||
from keras_core.layers.merging.base_merge import Merge
|
||||
|
||||
|
||||
@keras_core_export("keras_core.layers.Average")
|
||||
class Average(Merge):
|
||||
"""Averages a list of inputs element-wise..
|
||||
|
||||
It takes as input a list of tensors, all of the same shape,
|
||||
and returns a single tensor (also of the same shape).
|
||||
|
||||
Examples:
|
||||
|
||||
>>> input_shape = (2, 3, 4)
|
||||
>>> x1 = np.random.rand(*input_shape)
|
||||
>>> x2 = np.random.rand(*input_shape)
|
||||
>>> y = keras_core.layers.Average()([x1, x2])
|
||||
|
||||
Usage in a Keras model:
|
||||
|
||||
>>> input1 = keras_core.layers.Input(shape=(16,))
|
||||
>>> x1 = keras_core.layers.Dense(8, activation='relu')(input1)
|
||||
>>> input2 = keras_core.layers.Input(shape=(32,))
|
||||
>>> x2 = keras_core.layers.Dense(8, activation='relu')(input2)
|
||||
>>> # equivalent to `y = keras_core.layers.average([x1, x2])`
|
||||
>>> y = keras_core.layers.Average()([x1, x2])
|
||||
>>> out = keras_core.layers.Dense(4)(y)
|
||||
>>> model = keras_core.models.Model(inputs=[input1, input2], outputs=out)
|
||||
|
||||
"""
|
||||
|
||||
def _merge_function(self, inputs):
|
||||
output = inputs[0]
|
||||
for i in range(1, len(inputs)):
|
||||
output += inputs[i]
|
||||
return output / len(inputs)
|
||||
|
||||
|
||||
@keras_core_export("keras_core.layers.average")
|
||||
def average(inputs, **kwargs):
|
||||
"""Functional interface to the `keras_core.layers.Average` layer.
|
||||
|
||||
Args:
|
||||
inputs: A list of input tensors , all of the same shape.
|
||||
**kwargs: Standard layer keyword arguments.
|
||||
|
||||
Returns:
|
||||
A tensor as the element-wise product of the inputs with the same
|
||||
shape as the inputs.
|
||||
|
||||
Examples:
|
||||
|
||||
>>> input_shape = (2, 3, 4)
|
||||
>>> x1 = np.random.rand(*input_shape)
|
||||
>>> x2 = np.random.rand(*input_shape)
|
||||
>>> y = keras_core.layers.average([x1, x2])
|
||||
|
||||
Usage in a Keras model:
|
||||
|
||||
>>> input1 = keras_core.layers.Input(shape=(16,))
|
||||
>>> x1 = keras_core.layers.Dense(8, activation='relu')(input1)
|
||||
>>> input2 = keras_core.layers.Input(shape=(32,))
|
||||
>>> x2 = keras_core.layers.Dense(8, activation='relu')(input2)
|
||||
>>> y = keras_core.layers.average([x1, x2])
|
||||
>>> out = keras_core.layers.Dense(4)(y)
|
||||
>>> model = keras_core.models.Model(inputs=[input1, input2], outputs=out)
|
||||
|
||||
"""
|
||||
return Average(**kwargs)(inputs)
|
||||
70
keras_core/layers/merging/maximum.py
Normal file
70
keras_core/layers/merging/maximum.py
Normal file
@ -0,0 +1,70 @@
|
||||
from keras_core import operations as ops
|
||||
from keras_core.api_export import keras_core_export
|
||||
from keras_core.layers.merging.base_merge import Merge
|
||||
|
||||
|
||||
@keras_core_export("keras_core.layers.Maximum")
|
||||
class Maximum(Merge):
|
||||
"""Computes element-wise maximum on a list of inputs.
|
||||
|
||||
It takes as input a list of tensors, all of the same shape,
|
||||
and returns a single tensor (also of the same shape).
|
||||
|
||||
Examples:
|
||||
|
||||
>>> input_shape = (2, 3, 4)
|
||||
>>> x1 = np.random.rand(*input_shape)
|
||||
>>> x2 = np.random.rand(*input_shape)
|
||||
>>> y = keras_core.layers.Maximum()([x1, x2])
|
||||
|
||||
Usage in a Keras model:
|
||||
|
||||
>>> input1 = keras_core.layers.Input(shape=(16,))
|
||||
>>> x1 = keras_core.layers.Dense(8, activation='relu')(input1)
|
||||
>>> input2 = keras_core.layers.Input(shape=(32,))
|
||||
>>> x2 = keras_core.layers.Dense(8, activation='relu')(input2)
|
||||
>>> # equivalent to `y = keras_core.layers.maximum([x1, x2])`
|
||||
>>> y = keras_core.layers.Maximum()([x1, x2])
|
||||
>>> out = keras_core.layers.Dense(4)(y)
|
||||
>>> model = keras_core.models.Model(inputs=[input1, input2], outputs=out)
|
||||
|
||||
"""
|
||||
|
||||
def _merge_function(self, inputs):
|
||||
output = inputs[0]
|
||||
for i in range(1, len(inputs)):
|
||||
output = ops.maximum(output, inputs[i])
|
||||
return output
|
||||
|
||||
|
||||
@keras_core_export("keras_core.layers.maximum")
|
||||
def maximum(inputs, **kwargs):
|
||||
"""Functional interface to the `keras_core.layers.Maximum` layer.
|
||||
|
||||
Args:
|
||||
inputs: A list of input tensors , all of the same shape.
|
||||
**kwargs: Standard layer keyword arguments.
|
||||
|
||||
Returns:
|
||||
A tensor as the element-wise product of the inputs with the same
|
||||
shape as the inputs.
|
||||
|
||||
Examples:
|
||||
|
||||
>>> input_shape = (2, 3, 4)
|
||||
>>> x1 = np.random.rand(*input_shape)
|
||||
>>> x2 = np.random.rand(*input_shape)
|
||||
>>> y = keras_core.layers.maximum([x1, x2])
|
||||
|
||||
Usage in a Keras model:
|
||||
|
||||
>>> input1 = keras_core.layers.Input(shape=(16,))
|
||||
>>> x1 = keras_core.layers.Dense(8, activation='relu')(input1)
|
||||
>>> input2 = keras_core.layers.Input(shape=(32,))
|
||||
>>> x2 = keras_core.layers.Dense(8, activation='relu')(input2)
|
||||
>>> y = keras_core.layers.maximum([x1, x2])
|
||||
>>> out = keras_core.layers.Dense(4)(y)
|
||||
>>> model = keras_core.models.Model(inputs=[input1, input2], outputs=out)
|
||||
|
||||
"""
|
||||
return Maximum(**kwargs)(inputs)
|
||||
@ -199,3 +199,367 @@ class MergingLayersTest(testing.TestCase):
|
||||
ValueError, "layer should be called on exactly 2 inputs"
|
||||
):
|
||||
layers.Subtract()([input_1])
|
||||
|
||||
def test_minimum_basic(self):
|
||||
self.run_layer_test(
|
||||
layers.Minimum,
|
||||
init_kwargs={},
|
||||
input_shape=[(2, 3), (2, 3)],
|
||||
expected_output_shape=(2, 3),
|
||||
expected_num_trainable_weights=0,
|
||||
expected_num_non_trainable_weights=0,
|
||||
expected_num_seed_generators=0,
|
||||
expected_num_losses=0,
|
||||
supports_masking=True,
|
||||
)
|
||||
|
||||
@pytest.mark.skipif(
|
||||
not backend.DYNAMIC_SHAPES_OK,
|
||||
reason="Backend does not support dynamic shapes.",
|
||||
)
|
||||
def test_minimum_correctness_dynamic(self):
|
||||
x1 = np.random.rand(2, 4, 5)
|
||||
x2 = np.random.rand(2, 4, 5)
|
||||
x3 = ops.minimum(x1, x2)
|
||||
|
||||
input_1 = layers.Input(shape=(4, 5))
|
||||
input_2 = layers.Input(shape=(4, 5))
|
||||
merge_layer = layers.Minimum()
|
||||
out = merge_layer([input_1, input_2])
|
||||
model = models.Model([input_1, input_2], out)
|
||||
res = model([x1, x2])
|
||||
|
||||
self.assertEqual(res.shape, (2, 4, 5))
|
||||
self.assertAllClose(res, x3, atol=1e-4)
|
||||
self.assertIsNone(
|
||||
merge_layer.compute_mask([input_1, input_2], [None, None])
|
||||
)
|
||||
self.assertTrue(
|
||||
np.all(
|
||||
merge_layer.compute_mask(
|
||||
[input_1, input_2],
|
||||
[backend.Variable(x1), backend.Variable(x2)],
|
||||
)
|
||||
)
|
||||
)
|
||||
|
||||
def test_minimum_correctness_static(self):
|
||||
batch_size = 2
|
||||
shape = (4, 5)
|
||||
x1 = np.random.rand(batch_size, *shape)
|
||||
x2 = np.random.rand(batch_size, *shape)
|
||||
x3 = ops.minimum(x1, x2)
|
||||
|
||||
input_1 = layers.Input(shape=shape, batch_size=batch_size)
|
||||
input_2 = layers.Input(shape=shape, batch_size=batch_size)
|
||||
merge_layer = layers.Minimum()
|
||||
out = merge_layer([input_1, input_2])
|
||||
model = models.Model([input_1, input_2], out)
|
||||
res = model([x1, x2])
|
||||
|
||||
self.assertEqual(res.shape, (batch_size, *shape))
|
||||
self.assertAllClose(res, x3, atol=1e-4)
|
||||
self.assertIsNone(
|
||||
merge_layer.compute_mask([input_1, input_2], [None, None])
|
||||
)
|
||||
self.assertTrue(
|
||||
np.all(
|
||||
merge_layer.compute_mask(
|
||||
[input_1, input_2],
|
||||
[backend.Variable(x1), backend.Variable(x2)],
|
||||
)
|
||||
)
|
||||
)
|
||||
|
||||
def test_minimum_errors(self):
|
||||
batch_size = 2
|
||||
shape = (4, 5)
|
||||
x1 = np.random.rand(batch_size, *shape)
|
||||
|
||||
input_1 = layers.Input(shape=shape, batch_size=batch_size)
|
||||
input_2 = layers.Input(shape=shape, batch_size=batch_size)
|
||||
merge_layer = layers.Minimum()
|
||||
|
||||
with self.assertRaisesRegex(ValueError, "`mask` should be a list."):
|
||||
merge_layer.compute_mask([input_1, input_2], x1)
|
||||
|
||||
with self.assertRaisesRegex(ValueError, "`inputs` should be a list."):
|
||||
merge_layer.compute_mask(input_1, [None, None])
|
||||
|
||||
with self.assertRaisesRegex(
|
||||
ValueError, " should have the same length."
|
||||
):
|
||||
merge_layer.compute_mask([input_1, input_2], [None])
|
||||
|
||||
def test_maximum_basic(self):
|
||||
self.run_layer_test(
|
||||
layers.Maximum,
|
||||
init_kwargs={},
|
||||
input_shape=[(2, 3), (2, 3)],
|
||||
expected_output_shape=(2, 3),
|
||||
expected_num_trainable_weights=0,
|
||||
expected_num_non_trainable_weights=0,
|
||||
expected_num_seed_generators=0,
|
||||
expected_num_losses=0,
|
||||
supports_masking=True,
|
||||
)
|
||||
|
||||
@pytest.mark.skipif(
|
||||
not backend.DYNAMIC_SHAPES_OK,
|
||||
reason="Backend does not support dynamic shapes.",
|
||||
)
|
||||
def test_maximum_correctness_dynamic(self):
|
||||
x1 = np.random.rand(2, 4, 5)
|
||||
x2 = np.random.rand(2, 4, 5)
|
||||
x3 = ops.maximum(x1, x2)
|
||||
|
||||
input_1 = layers.Input(shape=(4, 5))
|
||||
input_2 = layers.Input(shape=(4, 5))
|
||||
merge_layer = layers.Maximum()
|
||||
out = merge_layer([input_1, input_2])
|
||||
model = models.Model([input_1, input_2], out)
|
||||
res = model([x1, x2])
|
||||
|
||||
self.assertEqual(res.shape, (2, 4, 5))
|
||||
self.assertAllClose(res, x3, atol=1e-4)
|
||||
self.assertIsNone(
|
||||
merge_layer.compute_mask([input_1, input_2], [None, None])
|
||||
)
|
||||
self.assertTrue(
|
||||
np.all(
|
||||
merge_layer.compute_mask(
|
||||
[input_1, input_2],
|
||||
[backend.Variable(x1), backend.Variable(x2)],
|
||||
)
|
||||
)
|
||||
)
|
||||
|
||||
def test_maximum_correctness_static(self):
|
||||
batch_size = 2
|
||||
shape = (4, 5)
|
||||
x1 = np.random.rand(batch_size, *shape)
|
||||
x2 = np.random.rand(batch_size, *shape)
|
||||
x3 = ops.maximum(x1, x2)
|
||||
|
||||
input_1 = layers.Input(shape=shape, batch_size=batch_size)
|
||||
input_2 = layers.Input(shape=shape, batch_size=batch_size)
|
||||
merge_layer = layers.Maximum()
|
||||
out = merge_layer([input_1, input_2])
|
||||
model = models.Model([input_1, input_2], out)
|
||||
res = model([x1, x2])
|
||||
|
||||
self.assertEqual(res.shape, (batch_size, *shape))
|
||||
self.assertAllClose(res, x3, atol=1e-4)
|
||||
self.assertIsNone(
|
||||
merge_layer.compute_mask([input_1, input_2], [None, None])
|
||||
)
|
||||
self.assertTrue(
|
||||
np.all(
|
||||
merge_layer.compute_mask(
|
||||
[input_1, input_2],
|
||||
[backend.Variable(x1), backend.Variable(x2)],
|
||||
)
|
||||
)
|
||||
)
|
||||
|
||||
def test_maximum_errors(self):
|
||||
batch_size = 2
|
||||
shape = (4, 5)
|
||||
x1 = np.random.rand(batch_size, *shape)
|
||||
|
||||
input_1 = layers.Input(shape=shape, batch_size=batch_size)
|
||||
input_2 = layers.Input(shape=shape, batch_size=batch_size)
|
||||
merge_layer = layers.Maximum()
|
||||
|
||||
with self.assertRaisesRegex(ValueError, "`mask` should be a list."):
|
||||
merge_layer.compute_mask([input_1, input_2], x1)
|
||||
|
||||
with self.assertRaisesRegex(ValueError, "`inputs` should be a list."):
|
||||
merge_layer.compute_mask(input_1, [None, None])
|
||||
|
||||
with self.assertRaisesRegex(
|
||||
ValueError, " should have the same length."
|
||||
):
|
||||
merge_layer.compute_mask([input_1, input_2], [None])
|
||||
|
||||
def test_multiply_basic(self):
|
||||
self.run_layer_test(
|
||||
layers.Multiply,
|
||||
init_kwargs={},
|
||||
input_shape=[(2, 3), (2, 3)],
|
||||
expected_output_shape=(2, 3),
|
||||
expected_num_trainable_weights=0,
|
||||
expected_num_non_trainable_weights=0,
|
||||
expected_num_seed_generators=0,
|
||||
expected_num_losses=0,
|
||||
supports_masking=True,
|
||||
)
|
||||
|
||||
@pytest.mark.skipif(
|
||||
not backend.DYNAMIC_SHAPES_OK,
|
||||
reason="Backend does not support dynamic shapes.",
|
||||
)
|
||||
def test_multiply_correctness_dynamic(self):
|
||||
x1 = np.random.rand(2, 4, 5)
|
||||
x2 = np.random.rand(2, 4, 5)
|
||||
x3 = ops.convert_to_tensor(x1 * x2)
|
||||
|
||||
input_1 = layers.Input(shape=(4, 5))
|
||||
input_2 = layers.Input(shape=(4, 5))
|
||||
merge_layer = layers.Multiply()
|
||||
out = merge_layer([input_1, input_2])
|
||||
model = models.Model([input_1, input_2], out)
|
||||
res = model([x1, x2])
|
||||
|
||||
self.assertEqual(res.shape, (2, 4, 5))
|
||||
self.assertAllClose(res, x3, atol=1e-4)
|
||||
self.assertIsNone(
|
||||
merge_layer.compute_mask([input_1, input_2], [None, None])
|
||||
)
|
||||
self.assertTrue(
|
||||
np.all(
|
||||
merge_layer.compute_mask(
|
||||
[input_1, input_2],
|
||||
[backend.Variable(x1), backend.Variable(x2)],
|
||||
)
|
||||
)
|
||||
)
|
||||
|
||||
def test_multiply_correctness_static(self):
|
||||
batch_size = 2
|
||||
shape = (4, 5)
|
||||
x1 = np.random.rand(batch_size, *shape)
|
||||
x2 = np.random.rand(batch_size, *shape)
|
||||
x3 = ops.convert_to_tensor(x1 * x2)
|
||||
|
||||
input_1 = layers.Input(shape=shape, batch_size=batch_size)
|
||||
input_2 = layers.Input(shape=shape, batch_size=batch_size)
|
||||
merge_layer = layers.Multiply()
|
||||
out = merge_layer([input_1, input_2])
|
||||
model = models.Model([input_1, input_2], out)
|
||||
res = model([x1, x2])
|
||||
|
||||
self.assertEqual(res.shape, (batch_size, *shape))
|
||||
self.assertAllClose(res, x3, atol=1e-4)
|
||||
self.assertIsNone(
|
||||
merge_layer.compute_mask([input_1, input_2], [None, None])
|
||||
)
|
||||
self.assertTrue(
|
||||
np.all(
|
||||
merge_layer.compute_mask(
|
||||
[input_1, input_2],
|
||||
[backend.Variable(x1), backend.Variable(x2)],
|
||||
)
|
||||
)
|
||||
)
|
||||
|
||||
def test_multiply_errors(self):
|
||||
batch_size = 2
|
||||
shape = (4, 5)
|
||||
x1 = np.random.rand(batch_size, *shape)
|
||||
|
||||
input_1 = layers.Input(shape=shape, batch_size=batch_size)
|
||||
input_2 = layers.Input(shape=shape, batch_size=batch_size)
|
||||
merge_layer = layers.Multiply()
|
||||
|
||||
with self.assertRaisesRegex(ValueError, "`mask` should be a list."):
|
||||
merge_layer.compute_mask([input_1, input_2], x1)
|
||||
|
||||
with self.assertRaisesRegex(ValueError, "`inputs` should be a list."):
|
||||
merge_layer.compute_mask(input_1, [None, None])
|
||||
|
||||
with self.assertRaisesRegex(
|
||||
ValueError, " should have the same length."
|
||||
):
|
||||
merge_layer.compute_mask([input_1, input_2], [None])
|
||||
|
||||
def test_average_basic(self):
|
||||
self.run_layer_test(
|
||||
layers.Average,
|
||||
init_kwargs={},
|
||||
input_shape=[(2, 3), (2, 3)],
|
||||
expected_output_shape=(2, 3),
|
||||
expected_num_trainable_weights=0,
|
||||
expected_num_non_trainable_weights=0,
|
||||
expected_num_seed_generators=0,
|
||||
expected_num_losses=0,
|
||||
supports_masking=True,
|
||||
)
|
||||
|
||||
@pytest.mark.skipif(
|
||||
not backend.DYNAMIC_SHAPES_OK,
|
||||
reason="Backend does not support dynamic shapes.",
|
||||
)
|
||||
def test_average_correctness_dynamic(self):
|
||||
x1 = np.random.rand(2, 4, 5)
|
||||
x2 = np.random.rand(2, 4, 5)
|
||||
x3 = ops.average(np.array([x1, x2]), axis=0)
|
||||
|
||||
input_1 = layers.Input(shape=(4, 5))
|
||||
input_2 = layers.Input(shape=(4, 5))
|
||||
merge_layer = layers.Average()
|
||||
out = merge_layer([input_1, input_2])
|
||||
model = models.Model([input_1, input_2], out)
|
||||
res = model([x1, x2])
|
||||
|
||||
self.assertEqual(res.shape, (2, 4, 5))
|
||||
self.assertAllClose(res, x3, atol=1e-4)
|
||||
self.assertIsNone(
|
||||
merge_layer.compute_mask([input_1, input_2], [None, None])
|
||||
)
|
||||
self.assertTrue(
|
||||
np.all(
|
||||
merge_layer.compute_mask(
|
||||
[input_1, input_2],
|
||||
[backend.Variable(x1), backend.Variable(x2)],
|
||||
)
|
||||
)
|
||||
)
|
||||
|
||||
def test_average_correctness_static(self):
|
||||
batch_size = 2
|
||||
shape = (4, 5)
|
||||
x1 = np.random.rand(batch_size, *shape)
|
||||
x2 = np.random.rand(batch_size, *shape)
|
||||
x3 = ops.average(np.array([x1, x2]), axis=0)
|
||||
|
||||
input_1 = layers.Input(shape=shape, batch_size=batch_size)
|
||||
input_2 = layers.Input(shape=shape, batch_size=batch_size)
|
||||
merge_layer = layers.Average()
|
||||
out = merge_layer([input_1, input_2])
|
||||
model = models.Model([input_1, input_2], out)
|
||||
res = model([x1, x2])
|
||||
|
||||
self.assertEqual(res.shape, (batch_size, *shape))
|
||||
self.assertAllClose(res, x3, atol=1e-4)
|
||||
self.assertIsNone(
|
||||
merge_layer.compute_mask([input_1, input_2], [None, None])
|
||||
)
|
||||
self.assertTrue(
|
||||
np.all(
|
||||
merge_layer.compute_mask(
|
||||
[input_1, input_2],
|
||||
[backend.Variable(x1), backend.Variable(x2)],
|
||||
)
|
||||
)
|
||||
)
|
||||
|
||||
def test_average_errors(self):
|
||||
batch_size = 2
|
||||
shape = (4, 5)
|
||||
x1 = np.random.rand(batch_size, *shape)
|
||||
|
||||
input_1 = layers.Input(shape=shape, batch_size=batch_size)
|
||||
input_2 = layers.Input(shape=shape, batch_size=batch_size)
|
||||
merge_layer = layers.Average()
|
||||
|
||||
with self.assertRaisesRegex(ValueError, "`mask` should be a list."):
|
||||
merge_layer.compute_mask([input_1, input_2], x1)
|
||||
|
||||
with self.assertRaisesRegex(ValueError, "`inputs` should be a list."):
|
||||
merge_layer.compute_mask(input_1, [None, None])
|
||||
|
||||
with self.assertRaisesRegex(
|
||||
ValueError, " should have the same length."
|
||||
):
|
||||
merge_layer.compute_mask([input_1, input_2], [None])
|
||||
|
||||
70
keras_core/layers/merging/minimum.py
Normal file
70
keras_core/layers/merging/minimum.py
Normal file
@ -0,0 +1,70 @@
|
||||
from keras_core import operations as ops
|
||||
from keras_core.api_export import keras_core_export
|
||||
from keras_core.layers.merging.base_merge import Merge
|
||||
|
||||
|
||||
@keras_core_export("keras_core.layers.Minimum")
|
||||
class Minimum(Merge):
|
||||
"""Computes elementwise minimum on a list of inputs.
|
||||
|
||||
It takes as input a list of tensors, all of the same shape,
|
||||
and returns a single tensor (also of the same shape).
|
||||
|
||||
Examples:
|
||||
|
||||
>>> input_shape = (2, 3, 4)
|
||||
>>> x1 = np.random.rand(*input_shape)
|
||||
>>> x2 = np.random.rand(*input_shape)
|
||||
>>> y = keras_core.layers.Minimum()([x1, x2])
|
||||
|
||||
Usage in a Keras model:
|
||||
|
||||
>>> input1 = keras_core.layers.Input(shape=(16,))
|
||||
>>> x1 = keras_core.layers.Dense(8, activation='relu')(input1)
|
||||
>>> input2 = keras_core.layers.Input(shape=(32,))
|
||||
>>> x2 = keras_core.layers.Dense(8, activation='relu')(input2)
|
||||
>>> # equivalent to `y = keras_core.layers.minimum([x1, x2])`
|
||||
>>> y = keras_core.layers.Minimum()([x1, x2])
|
||||
>>> out = keras_core.layers.Dense(4)(y)
|
||||
>>> model = keras_core.models.Model(inputs=[input1, input2], outputs=out)
|
||||
|
||||
"""
|
||||
|
||||
def _merge_function(self, inputs):
|
||||
output = inputs[0]
|
||||
for i in range(1, len(inputs)):
|
||||
output = ops.minimum(output, inputs[i])
|
||||
return output
|
||||
|
||||
|
||||
@keras_core_export("keras_core.layers.minimum")
|
||||
def minimum(inputs, **kwargs):
|
||||
"""Functional interface to the `keras_core.layers.Minimum` layer.
|
||||
|
||||
Args:
|
||||
inputs: A list of input tensors , all of the same shape.
|
||||
**kwargs: Standard layer keyword arguments.
|
||||
|
||||
Returns:
|
||||
A tensor as the elementwise product of the inputs with the same
|
||||
shape as the inputs.
|
||||
|
||||
Examples:
|
||||
|
||||
>>> input_shape = (2, 3, 4)
|
||||
>>> x1 = np.random.rand(*input_shape)
|
||||
>>> x2 = np.random.rand(*input_shape)
|
||||
>>> y = keras_core.layers.minimum([x1, x2])
|
||||
|
||||
Usage in a Keras model:
|
||||
|
||||
>>> input1 = keras_core.layers.Input(shape=(16,))
|
||||
>>> x1 = keras_core.layers.Dense(8, activation='relu')(input1)
|
||||
>>> input2 = keras_core.layers.Input(shape=(32,))
|
||||
>>> x2 = keras_core.layers.Dense(8, activation='relu')(input2)
|
||||
>>> y = keras_core.layers.minimum([x1, x2])
|
||||
>>> out = keras_core.layers.Dense(4)(y)
|
||||
>>> model = keras_core.models.Model(inputs=[input1, input2], outputs=out)
|
||||
|
||||
"""
|
||||
return Minimum(**kwargs)(inputs)
|
||||
69
keras_core/layers/merging/multiply.py
Normal file
69
keras_core/layers/merging/multiply.py
Normal file
@ -0,0 +1,69 @@
|
||||
from keras_core.api_export import keras_core_export
|
||||
from keras_core.layers.merging.base_merge import Merge
|
||||
|
||||
|
||||
@keras_core_export("keras_core.layers.Multiply")
|
||||
class Multiply(Merge):
|
||||
"""Performs elementwise multiplication.
|
||||
|
||||
It takes as input a list of tensors, all of the same shape,
|
||||
and returns a single tensor (also of the same shape).
|
||||
|
||||
Examples:
|
||||
|
||||
>>> input_shape = (2, 3, 4)
|
||||
>>> x1 = np.random.rand(*input_shape)
|
||||
>>> x2 = np.random.rand(*input_shape)
|
||||
>>> y = keras_core.layers.Multiply()([x1, x2])
|
||||
|
||||
Usage in a Keras model:
|
||||
|
||||
>>> input1 = keras_core.layers.Input(shape=(16,))
|
||||
>>> x1 = keras_core.layers.Dense(8, activation='relu')(input1)
|
||||
>>> input2 = keras_core.layers.Input(shape=(32,))
|
||||
>>> x2 = keras_core.layers.Dense(8, activation='relu')(input2)
|
||||
>>> # equivalent to `y = keras_core.layers.multiply([x1, x2])`
|
||||
>>> y = keras_core.layers.Multiply()([x1, x2])
|
||||
>>> out = keras_core.layers.Dense(4)(y)
|
||||
>>> model = keras_core.models.Model(inputs=[input1, input2], outputs=out)
|
||||
|
||||
"""
|
||||
|
||||
def _merge_function(self, inputs):
|
||||
output = inputs[0]
|
||||
for i in range(1, len(inputs)):
|
||||
output = output * inputs[i]
|
||||
return output
|
||||
|
||||
|
||||
@keras_core_export("keras_core.layers.multiply")
|
||||
def multiply(inputs, **kwargs):
|
||||
"""Functional interface to the `keras_core.layers.Multiply` layer.
|
||||
|
||||
Args:
|
||||
inputs: A list of input tensors , all of the same shape.
|
||||
**kwargs: Standard layer keyword arguments.
|
||||
|
||||
Returns:
|
||||
A tensor as the elementwise product of the inputs with the same
|
||||
shape as the inputs.
|
||||
|
||||
Examples:
|
||||
|
||||
>>> input_shape = (2, 3, 4)
|
||||
>>> x1 = np.random.rand(*input_shape)
|
||||
>>> x2 = np.random.rand(*input_shape)
|
||||
>>> y = keras_core.layers.multiply([x1, x2])
|
||||
|
||||
Usage in a Keras model:
|
||||
|
||||
>>> input1 = keras_core.layers.Input(shape=(16,))
|
||||
>>> x1 = keras_core.layers.Dense(8, activation='relu')(input1)
|
||||
>>> input2 = keras_core.layers.Input(shape=(32,))
|
||||
>>> x2 = keras_core.layers.Dense(8, activation='relu')(input2)
|
||||
>>> y = keras_core.layers.multiply([x1, x2])
|
||||
>>> out = keras_core.layers.Dense(4)(y)
|
||||
>>> model = keras_core.models.Model(inputs=[input1, input2], outputs=out)
|
||||
|
||||
"""
|
||||
return Multiply(**kwargs)(inputs)
|
||||
Loading…
Reference in New Issue
Block a user