Cumulative Update for Keras 2.2.0 and some fixes and Explanations in the README.md

README.md

@@ -24,8 +24,8 @@ from keras.layers import *
 from keras.models import *
 from recurrentshop import *
 
-x_t = Input(5,)) # The input to the RNN at time t
-h_tm1 = Input((10,))  # Previous hidden state
+x_t = Input(shape=(5,)) # The input to the RNN at time t
+h_tm1 = Input(shape=(10,))  # Previous hidden state
 
 # Compute new hidden state
 h_t = add([Dense(10)(x_t), Dense(10, use_bias=False)(h_tm1)])
@@ -34,17 +34,35 @@ h_t = add([Dense(10)(x_t), Dense(10, use_bias=False)(h_tm1)])
 h_t = Activation('tanh')(h_t)
 
 # Build the RNN
-rnn = RecurrentModel(input=x_t, initial_states=[h_tm1], output=h_t, output_states=[h_t])
-
-# rnn is a standard Keras `Recurrent` instance. RecuurentModel also accepts arguments such as unroll, return_sequences etc
+# RecurrentModel is a standard Keras `Recurrent` layer. 
+# RecurrentModel also accepts arguments such as unroll, return_sequences etc
+rnn = RecurrentModel(input=x_t, initial_states=[h_tm1], output=h_t, final_states=[h_t])
+# return_sequences is False by default
+# so it only returns the last h_t state
+
+# Build a Keras Model using our RNN layer
+# input dimensions are (Time_steps, Depth)
+x = Input(shape=(7,5))
+y = rnn(x)
+model = Model(x, y)
 
 # Run the RNN over a random sequence
+# Don't forget the batch shape when calling the model!
+out=model.predict(np.random.random((1, 7, 5)))
+print(out.shape)#->(1,10)
 
-x = Input((7,5))
-y = rnn(x)
 
-model = Model(x, y)
-model.predict(np.random.random((7, 5)))
+# to get one output per input sequence element, set return_sequences=True
+rnn2 = RecurrentModel(input=x_t, initial_states=[h_tm1], output=h_t, final_states=[h_t],return_sequences=True)
+
+# Time_steps can also be None to allow variable Sequence Length
+# Note that this is not compatible with unroll=True
+x = Input(shape=(None ,5))
+y = rnn2(x)
+model2 = Model(x, y)
+
+out2=model2.predict(np.random.random((1, 7, 5)))
+print(out2.shape)#->(1,7,10)
 
 ```
 
@@ -129,7 +147,7 @@ See docs/ directory for more features.
 # Installation
 
 ```shell
-git clone https://www.github.com/datalogai/recurrentshop.git
+git clone https://www.github.com/farizrahman4u/recurrentshop.git
 cd recurrentshop
 python setup.py install
 ```

recurrentshop/engine.py

@@ -3,10 +3,13 @@
 from keras import initializers
 from .backend import rnn, learning_phase_scope
 from .generic_utils import serialize_function, deserialize_function
-from keras.engine.topology import Node, _collect_previous_mask, _collect_input_shape
+from keras.engine.base_layer import Node,_collect_previous_mask, _collect_input_shape
 import inspect
 
 
+if K.backend() == 'tensorflow':
+    import tensorflow as tf
+
 def _to_list(x):
     if type(x) is not list:
         x = [x]
@@ -549,9 +552,12 @@ def call(self, inputs, initial_state=None, initial_readout=None, ground_truth=No
             if self.teacher_force:
                 if ground_truth is None or self._is_optional_input_placeholder(ground_truth):
                     raise Exception('ground_truth must be provided for RecurrentModel with teacher_force=True.')
-                # counter = K.zeros((1,), dtype='int32')
-                counter = K.zeros((1,))
-                counter = K.cast(counter, 'int32')
+                if K.backend() == 'tensorflow':
+                    with tf.control_dependencies(None):
+                        counter = K.zeros((1,))
+                else:
+                    counter = K.zeros((1,))
+                counter = K.cast(counter, 'int32')  
                 initial_states.insert(-1, counter)
                 initial_states[-2]
                 initial_states.insert(-1, ground_truth)
@@ -654,8 +660,13 @@ def step(self, inputs, states):
             ground_truth = states.pop()
             assert K.ndim(ground_truth) == 3, K.ndim(ground_truth)
             counter = states.pop()
-            zero = K.cast(K.zeros((1,))[0], 'int32')
-            one = K.cast(K.zeros((1,))[0], 'int32')
+            if K.backend() == 'tensorflow':
+                with tf.control_dependencies(None):
+                    zero = K.cast(K.zeros((1,))[0], 'int32')
+                    one = K.cast(K.zeros((1,))[0], 'int32')
+            else:
+                zero = K.cast(K.zeros((1,))[0], 'int32')
+                one = K.cast(K.zeros((1,))[0], 'int32')
             slices = [slice(None), counter[0] - K.switch(counter[0], one, zero)] + [slice(None)] * (K.ndim(ground_truth) - 2)
             ground_truth_slice = ground_truth[slices]
             readout = K.in_train_phase(K.switch(counter[0], ground_truth_slice, readout), readout)
@@ -839,13 +850,13 @@ def _get_optional_input_placeholder(self, name=None, num=1):
                     self._optional_input_placeholders[name] = self._get_optional_input_placeholder()
             return self._optional_input_placeholders[name]
         if num == 1:
-            optional_input_placeholder = _to_list(_OptionalInputPlaceHolder().inbound_nodes[0].output_tensors)[0]
+            optional_input_placeholder = _to_list(_OptionalInputPlaceHolder()._inbound_nodes[0].output_tensors)[0]
             assert self._is_optional_input_placeholder(optional_input_placeholder)
             return optional_input_placeholder
         else:
             y = []
             for _ in range(num):
-                optional_input_placeholder = _to_list(_OptionalInputPlaceHolder().inbound_nodes[0].output_tensors)[0]
+                optional_input_placeholder = _to_list(_OptionalInputPlaceHolder()._inbound_nodes[0].output_tensors)[0]
                 assert self._is_optional_input_placeholder(optional_input_placeholder)
                 y.append(optional_input_placeholder)
             return y

requirements.txt

@@ -1,3 +1,3 @@
-keras>=2.0.0
+keras>=2.2.0
 numpy>=1.8.0
-theano>=0.8.0
+tensorflow>=1.9