metrics.py

import tensorflow as tf
import scipy.io as sio


def masked_softmax_cross_entropy(preds, labels, mask):
    """Softmax cross-entropy loss with masking."""
    loss = tf.nn.softmax_cross_entropy_with_logits(logits=preds, labels=labels)
    mask = tf.cast(mask, dtype=tf.float32)
    mask /= tf.reduce_mean(mask)
    loss *= mask
    return tf.reduce_mean(loss)


def masked_accuracy(preds, labels, mask):
    """Accuracy with masking."""
    correct_prediction = tf.equal(tf.argmax(preds, 1), tf.argmax(labels, 1))
    accuracy_all = tf.cast(correct_prediction, tf.float32)
    mask = tf.cast(mask, dtype=tf.float32)
    mask /= tf.reduce_mean(mask)
    accuracy_all *= mask
    return tf.reduce_mean(accuracy_all)


def predictions(preds):
    """return prediction"""
    
    prediction = tf.argmax(preds, 1)
   # prediction = tf.Print(prediction, [prediction])
    return prediction