Created by: guru4elephant

add recompute based checkpoints methods for large batch training test=develop

example for usage

import paddle.fluid as fluid

def mlp(input_x, input_y, hid_dim=128, label_dim=2):
    fc_1 = fluid.layers.fc(input=input_x, size=hid_dim)
    prediction = fluid.layers.fc(input=[fc_1], size=label_dim, act='softmax')
    cost = fluid.layers.cross_entropy(input=prediction, label=input_y)
    sum_cost = fluid.layers.reduce_sum(cost)
    return sum_cost, fc_1, prediction

input_x = fluid.layers.data(name="x", shape=[32], dtype='float32')
input_y = fluid.layers.data(name="y", shape=[1], dtype='int64')

cost, fc_1, pred = mlp(input_x, input_y)

sgd = fluid.optimizer.SGD(learning_rate=0.01)
sgd = fluid.optimizer.RecomputeOptimizer(sgd)
sgd._set_checkpoints([fc_1, pred])
sgd.minimize(cost)