Update Sequential model to raise error for multi-output layer in the deferred mode.
This is a update for #36624, which we should show explicit error, rather than let the code proceed and failed down the road.
PiperOrigin-RevId: 295248095
Change-Id: I9cf9d0267f222c3994f3199bd9b49d86196ebb3b
diff --git a/tensorflow/python/keras/engine/sequential.py b/tensorflow/python/keras/engine/sequential.py
index 8609449..a86084f 100644
--- a/tensorflow/python/keras/engine/sequential.py
+++ b/tensorflow/python/keras/engine/sequential.py
@@ -39,6 +39,11 @@
from tensorflow.python.util.tf_export import keras_export
+SINGLE_LAYER_OUTPUT_ERROR_MSG = ('All layers in a Sequential model should have '
+ 'a single output tensor. For multi-output '
+ 'layers, use the functional API.')
+
+
@keras_export('keras.Sequential', 'keras.models.Sequential')
class Sequential(training.Model):
"""`Sequential` groups a linear stack of layers into a `tf.keras.Model`.
@@ -195,10 +200,7 @@
if set_inputs:
# If an input layer (placeholder) is available.
if len(nest.flatten(layer._inbound_nodes[-1].output_tensors)) != 1:
- raise ValueError('All layers in a Sequential model '
- 'should have a single output tensor. '
- 'For multi-output layers, '
- 'use the functional API.')
+ raise ValueError(SINGLE_LAYER_OUTPUT_ERROR_MSG)
self.outputs = [
nest.flatten(layer._inbound_nodes[-1].output_tensors)[0]
]
@@ -209,10 +211,7 @@
# refresh its output.
output_tensor = layer(self.outputs[0])
if len(nest.flatten(output_tensor)) != 1:
- raise TypeError('All layers in a Sequential model '
- 'should have a single output tensor. '
- 'For multi-output layers, '
- 'use the functional API.')
+ raise ValueError(SINGLE_LAYER_OUTPUT_ERROR_MSG)
self.outputs = [output_tensor]
if self.outputs:
@@ -286,6 +285,8 @@
outputs = layer(inputs, **kwargs)
+ if len(nest.flatten(outputs)) != 1:
+ raise ValueError(SINGLE_LAYER_OUTPUT_ERROR_MSG)
# `outputs` will be the inputs to the next layer.
inputs = outputs
mask = outputs._keras_mask
diff --git a/tensorflow/python/keras/engine/sequential_test.py b/tensorflow/python/keras/engine/sequential_test.py
index fa516eb..65e58fd 100644
--- a/tensorflow/python/keras/engine/sequential_test.py
+++ b/tensorflow/python/keras/engine/sequential_test.py
@@ -215,23 +215,6 @@
model = keras.models.Sequential()
model.add(None)
- # Added layers cannot have multiple outputs
- class MyLayer(keras.layers.Layer):
-
- def call(self, inputs):
- return [3 * inputs, 2 * inputs]
-
- def compute_output_shape(self, input_shape):
- return [input_shape, input_shape]
-
- with self.assertRaises(ValueError):
- model = keras.models.Sequential()
- model.add(MyLayer(input_shape=(3,)))
- with self.assertRaises(TypeError):
- model = keras.models.Sequential()
- model.add(keras.layers.Dense(1, input_dim=1))
- model.add(MyLayer())
-
@keras_parameterized.run_all_keras_modes
def test_nested_sequential_trainability(self):
input_dim = 20
@@ -405,6 +388,17 @@
ValueError, 'should have a single output tensor'):
keras.Sequential([MultiOutputLayer(input_shape=(3,))])
+ with self.assertRaisesRegexp(
+ ValueError, 'should have a single output tensor'):
+ keras.Sequential([
+ keras.layers.Dense(1, input_shape=(3,)),
+ MultiOutputLayer()])
+
+ # Should also raise error in a deferred build mode
+ with self.assertRaisesRegexp(
+ ValueError, 'should have a single output tensor'):
+ keras.Sequential([MultiOutputLayer()])(np.zeros((10, 10)))
+
@keras_parameterized.run_all_keras_modes
def test_layer_add_after_compile_deferred(self):
model = keras.Sequential([keras.layers.Dense(3)])
