b688192cbd74495c9907b6dcfce47e9a582ae1af,keras/backend/tensorflow_backend.py,,rnn,#Any#Any#Any#Any#Any#,391

Before Change

        // tensorflow yet for some reason it fails to in some test-cases. This
        // fixes the issue, but should be removed in future.
        mask.set_shape([inputs_shape[0].value, inputs_shape[1].value, 1])
        mask = tf.cast(mask, tf.bool)
    else:
        // Transpose not supported by bool tensor types, hence round-trip to uint8.
        mask = tf.cast(tf.transpose(tf.cast(mask, tf.uint8), axes), tf.bool)

After Change

    if go_backwards:
        input_list.reverse()

    if mask is not None:
        // Transpose not supported by bool tensor types, hence round-trip to uint8.
        mask = tf.cast(tf.transpose(tf.cast(mask, tf.uint8), axes), tf.bool)
        mask_list = tf.unpack(mask)

        for input, mask_t in zip(input_list, mask_list):
            output, new_states = step_function(input, states)

            // tf.select needs its condition tensor to be the same shape as its two
            // result tensors, but in our case the condition (mask) tensor is
            // (nsamples, 1), and A and B are (nsamples, ndimensions). So we need to
            // broadcast the mask to match the shape of A and B. That"s what the
            // tile call does, is just repeat the mask along its second dimension
            // ndimensions times.
            tiled_mask_t = tf.tile(mask_t, tf.pack([1, tf.shape(output)[1]]))

            if len(successive_outputs) == 0:
                prev_output = zeros_like(output)
            else:
                prev_output = successive_outputs[-1]

            output = tf.select(tiled_mask_t, output, prev_output)

            return_states = []
            for state, new_state in zip(states, new_states):
                // (see earlier comment for tile explanation)
                tiled_mask_t = tf.tile(mask_t, tf.pack([1, tf.shape(new_state)[1]]))
                return_states.append(tf.select(tiled_mask_t, new_state, state))

            states = return_states
            successive_outputs.append(output)
            successive_states.append(states)
    else:
        for input in input_list:
            output, states = step_function(input, states)
            successive_outputs.append(output)
            successive_states.append(states)

    last_output = successive_outputs[-1]
    outputs = tf.pack(successive_outputs)
    new_states = successive_states[-1]

Instances