mytheano_utils.py 文件源码

def remove_adjdup(x):
    """
    Remove adjacent duplicate items of a vector
    x: vector
    return a vector with adjacent duplicate items removed, for example [1,2,2,2,3,3,4] -> [1,2,3,4]
    """
    def update(x, nondup, idx):
        nondup = tensor.switch(tensor.eq(nondup[idx], x), nondup, tensor.set_subtensor(nondup[idx + 1], x))  # tensor.switch is much faster than ifelse
        idx = tensor.switch(tensor.eq(nondup[idx], x), idx, idx + 1)
        return nondup, idx
    nondup = x
    idx = tensor.as_tensor_variable(0)
    idx = tensor.cast(idx, 'int32')
    result, updates = theano.scan(fn = update, sequences=x, outputs_info=[nondup, idx], name='remove_adjdup')
    nondup = result[0][-1]
    idx = result[1][-1]
    return nondup[0:idx+1]