Performance issue in the definition of prepare_master_network, agent.py(P1)

Hello, I found a performance issue in the definition of `prepare_master_network`, agent.py, [`tf.reduce_sum(tf.square(pi.vf - r))`](https://github.com/carpedm20/SPIRAL-tensorflow/blob/2cca458f5d0d856c7bb73e877eea24906eda522f/agent.py#L186) will be created repeatedly during program execution, resulting in reduced efficiency. I think it should be created before the loop.

Looking forward to your reply. Btw, I am very glad to create a PR to fix it if you are too busy.