machine learning - running trained convolutional neural network -

so here question. have trained convolutional neural network classify images 2 classes using tensorflow. wondering how use weights neural network , test on unlabeled random image. there function in tensorflow or should run convolution on own now?

after finished training, can create

feed_dict_unlabeled = {x: x_unlabeled} 

now, using y_pred_cls defined in code follows:

y_pred_cls = tf.argmax(y_pred, dimension=1) 

you can do

y_labels = session.run(y_pred_cls, feed_dict=feed_dict_unlabeled)

to find labels of unlabeled data.

in addition, here discussion similar scenario may find helpful: python/tensorflow - have trained convolutional neural network, how test it?