Restore seed of weights in tensorflow

I trained a multi layer perceptron with tensorflow and initialized the weights like this:

weights = tf.get_variable(some_name, some_shape, initializer=tf.contrib.layers.variance_scaling_initializer())

After training I saved the graph on my disk:

saver = tf.train.Saver()
saver.save(sess, save_folder + filename)

My problem is, that I didnt set a seed for the weights. I want to retrain the model and get the exact same results. Is there any way whatsoever to get the random seed that was generated? Maybe out of the saved computation graph somehow or is there any other way to obtain the same results? I'm grateful for every suggestion.