Merge pull request #262 from philipperemy/fix_256

Fix bug introduced in Keras 3

.github/workflows/ci.yml

@@ -9,7 +9,7 @@ jobs:
  strategy:
  max-parallel: 4
  matrix:
- python-version: [ 3.9 ]
+ python-version: [ "3.10" ]
 
  steps:
  - uses: actions/checkout@v4

tasks/multi_length_sequences.py

@@ -46,4 +46,4 @@ def get_x_y(max_time_steps):
 m.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])
 
 gen = get_x_y(max_time_steps=MAX_TIME_STEP)
-m.fit(gen, epochs=1, steps_per_epoch=1000, max_queue_size=1, verbose=2)
+m.fit(gen, epochs=1, steps_per_epoch=1000, verbose=2)

tasks/plot_tcn_model.py

@@ -8,9 +8,9 @@
 num_features = 4
 
 inputs = tf.keras.layers.Input(shape=input_shape, name='input')
-tcn_out = TCN(nb_filters=64, kernel_size=3, nb_stacks=1, activation='LeakyReLU')(inputs)
+tcn_out = TCN(nb_filters=64, kernel_size=3, nb_stacks=1, activation='relu')(inputs)
 outputs = tf.keras.layers.Dense(forecast_horizon * num_features, activation='linear')(tcn_out)
-outputs = tf.reshape(outputs, shape=(-1, forecast_horizon, num_features), name='ouput')
+outputs = tf.keras.layers.Reshape((forecast_horizon, num_features), name='ouput')(outputs)
 model = tf.keras.Model(inputs=inputs, outputs=outputs)
 
 tf.keras.utils.plot_model(

tasks/save_reload_sequential_model.py

@@ -21,7 +21,7 @@
 with open('model.json', "w") as json_file:
  json_file.write(model_as_json)
 # save weights to file (for this format, need h5py installed)
-model.save_weights('weights.h5')
+model.save_weights('model.weights.h5')
 
 # Make inference.
 inputs = np.ones(shape=(1, 100))
@@ -36,7 +36,7 @@
 tcn_full_summary(model, expand_residual_blocks=False)
 
 # restore weights
-reloaded_model.load_weights('weights.h5')
+reloaded_model.load_weights('model.weights.h5')
 
 # Make inference.
 out2 = reloaded_model.predict(inputs)[0, 0]

tasks/tcn_call_test.py

@@ -3,7 +3,6 @@
 import numpy as np
 from tensorflow.keras import Input
 from tensorflow.keras import Model
-from tensorflow.keras.models import Sequential
 
 from tcn import TCN
 
@@ -100,27 +99,6 @@ def test_non_causal_time_dim_unknown_return_no_sequences(self):
  r = predict_with_tcn(time_steps=None, padding='same', return_sequences=False)
  self.assertListEqual([list(b.shape) for b in r], [[1, NB_FILTERS], [1, NB_FILTERS], [1, NB_FILTERS]])
 
- def test_norms(self):
- Sequential(layers=[TCN(input_shape=(20, 2), use_weight_norm=True)]).compile(optimizer='adam', loss='mse')
- Sequential(layers=[TCN(input_shape=(20, 2), use_weight_norm=False)]).compile(optimizer='adam', loss='mse')
- Sequential(layers=[TCN(input_shape=(20, 2), use_layer_norm=True)]).compile(optimizer='adam', loss='mse')
- Sequential(layers=[TCN(input_shape=(20, 2), use_layer_norm=False)]).compile(optimizer='adam', loss='mse')
- Sequential(layers=[TCN(input_shape=(20, 2), use_batch_norm=True)]).compile(optimizer='adam', loss='mse')
- Sequential(layers=[TCN(input_shape=(20, 2), use_batch_norm=False)]).compile(optimizer='adam', loss='mse')
- try:
- Sequential(layers=[TCN(input_shape=(20, 2), use_batch_norm=True, use_weight_norm=True)]).compile(
- optimizer='adam', loss='mse')
- raise AssertionError('test failed.')
- except ValueError:
- pass
- try:
- Sequential(layers=[TCN(input_shape=(20, 2), use_batch_norm=True,
- use_weight_norm=True, use_layer_norm=True)]).compile(
- optimizer='adam', loss='mse')
- raise AssertionError('test failed.')
- except ValueError:
- pass
-
  def test_receptive_field(self):
  self.assertEqual(37, TCN(kernel_size=3, dilations=(1, 3, 5), nb_stacks=1).receptive_field)
  self.assertEqual(379, TCN(kernel_size=4, dilations=(1, 2, 4, 8, 16, 32), nb_stacks=1).receptive_field)

tcn/tcn.py

@@ -1,5 +1,5 @@
 import inspect
-from typing import List # noqa
+from typing import List  # noqa
 
 import tensorflow as tf
 # pylint: disable=E0611,E0401
@@ -270,6 +270,12 @@ def __init__(self,
  def receptive_field(self):
  return 1 + 2 * (self.kernel_size - 1) * self.nb_stacks * sum(self.dilations)
 
+ def tolist(self, shape):
+ try:
+ return shape.as_list()
+ except AttributeError:
+ return shape
+
  def build(self, input_shape):
 
  # member to hold current output shape of the layer for building purposes
@@ -305,17 +311,17 @@ def build(self, input_shape):
 
  self.output_slice_index = None
  if self.padding == 'same':
- time = self.build_output_shape.as_list()[1]
+ time = self.tolist(self.build_output_shape)[1]
  if time is not None: # if time dimension is defined. e.g. shape = (bs, 500, input_dim).
- self.output_slice_index = int(self.build_output_shape.as_list()[1] / 2)
+ self.output_slice_index = int(self.tolist(self.build_output_shape)[1] / 2)
  else:
  # It will known at call time. c.f. self.call.
  self.padding_same_and_time_dim_unknown = True
 
  else:
  self.output_slice_index = -1 # causal case.
  self.slicer_layer = Lambda(lambda tt: tt[:, self.output_slice_index, :], name='Slice_Output')
- self.slicer_layer.build(self.build_output_shape.as_list())
+ self.slicer_layer.build(self.tolist(self.build_output_shape))
 
  def compute_output_shape(self, input_shape):
  """

tox.ini

@@ -1,5 +1,5 @@
 [tox]
-envlist = {py3}-tensorflow-{2.9,2.10,2.11,2.12,2.13,2.14,2.15}
+envlist = {py3}-tensorflow-{2.13,2.16,2.17}
 
 [testenv]
 setenv =
@@ -8,13 +8,9 @@ deps = pytest
  pylint
  flake8
  -rrequirements.txt
- tensorflow-2.9: tensorflow==2.9
- tensorflow-2.10: tensorflow==2.10
- tensorflow-2.11: tensorflow==2.11
- tensorflow-2.12: tensorflow==2.12
  tensorflow-2.13: tensorflow==2.13
- tensorflow-2.14: tensorflow==2.14
- tensorflow-2.15: tensorflow==2.15
+ tensorflow-2.16: tensorflow==2.16.2
+ tensorflow-2.17: tensorflow==2.17
 changedir = tasks/
 commands = pylint --disable=R,C,W,E1136 ../tcn
  flake8 ../tcn --count --select=E9,F63,F7,F82 --show-source --statistics