Remove skip_connnection option from ConvLSTM.

PiperOrigin-RevId: 192131304

sonnet/python/modules/gated_rnn.py

@@ -1236,7 +1236,6 @@ def __init__(self,
                rate=1,
                padding=conv.SAME,
                use_bias=True,
-               skip_connection=False,
                forget_bias=1.0,
                initializers=None,
                partitioners=None,
@@ -1259,8 +1258,6 @@ def __init__(self,
           convolution. Cannot be > 1 if any of stride is also > 1.
       padding: Padding algorithm, either `snt.SAME` or `snt.VALID`.
       use_bias: Use bias in convolutions.
-      skip_connection: If set to `True`, concatenate the input to the output
-          of the conv LSTM. Default: `False`.
       forget_bias: Forget bias.
       initializers: Dict containing ops to initialize the convolutional weights.
       partitioners: Optional dict containing partitioners to partition
@@ -1281,9 +1278,6 @@ def __init__(self,
 
     self._conv_class = self._get_conv_class(conv_ndims)
 
-    if skip_connection and stride != 1:
-      raise ValueError("`stride` needs to be 1 when using skip connection")
-
     if conv_ndims != len(input_shape)-1:
       raise ValueError("Invalid input_shape {} for conv_ndims={}.".format(
           input_shape, conv_ndims))
@@ -1297,16 +1291,13 @@ def __init__(self,
     self._padding = padding
     self._use_bias = use_bias
     self._forget_bias = forget_bias
-    self._skip_connection = skip_connection
     self._initializers = initializers
     self._partitioners = partitioners
     self._regularizers = regularizers
 
     self._total_output_channels = output_channels
     if self._stride != 1:
       self._total_output_channels //= self._stride * self._stride
-    if self._skip_connection:
-      self._total_output_channels += self._input_shape[-1]
 
     self._convolutions = collections.defaultdict(self._new_convolution)
 
@@ -1352,9 +1343,6 @@ def _build(self, inputs, state):
     next_cell = tf.sigmoid(forget_gate + self._forget_bias) * cell
     next_cell += tf.sigmoid(input_gate) * tf.tanh(next_input)
     output = tf.tanh(next_cell) * tf.sigmoid(output_gate)
-
-    if self._skip_connection:
-      output = tf.concat([output, inputs], axis=-1)
     return output, (output, next_cell)