How to implement gpu memory resource adaptation by tfgo?

[fcgxz2003]

When I used tf.LoadSavedModel to load tensorflow models saved in python and used Session.Run to perform training, I found that gpu memory allocation was very high. But I didn't have this problem when I ran the training in python. I guess it's because tensorflow pre-allocates gpu memory in tfgo.

my question is
My question is how should I disable pre-allocation in tfgo to achieve the following effect in python？

config = tf.ConfigProto()  
config.gpu_options.allow_growth=True  
sess = tf.Session(config=config)

Or setting like tf.ConfigProto()