Update Mo.py

Mo.py

@@ -226,13 +226,6 @@ def smoothed_loss_keras(y_true, y_pred, eps=0.1):
     per_tok = per_tok * mask
     return tf.reduce_sum(per_tok) / (tf.reduce_sum(mask) + 1e-8)
 
-def masked_accuracy(y_true, y_pred):
-    y_true = tf.cast(y_true, tf.int32)
-    mask = tf.cast(tf.not_equal(y_true, pad_id), tf.float32)
-    pred_id = tf.argmax(y_pred, axis=-1, output_type=tf.int32)
-    acc = tf.cast(tf.equal(y_true, pred_id), tf.float32) * mask
-    return tf.reduce_sum(acc) / (tf.reduce_sum(mask) + 1e-8)
-
 def masked_perplexity(y_true, y_pred, eps=0.1):
     y_true = tf.cast(y_true, tf.int32)
     mask = tf.cast(tf.not_equal(y_true, pad_id), tf.float32)
@@ -255,7 +248,7 @@ with strategy.scope():
     model.summary()
 
     optimizer = tf.keras.optimizers.Adam(1e-4, beta_1=0.9, beta_2=0.95, epsilon=1e-8, clipnorm=1.0)
-    model.compile(optimizer=optimizer, loss=smoothed_loss_keras, metrics=[masked_accuracy, masked_perplexity])
+    model.compile(optimizer=optimizer, loss=smoothed_loss_keras, metrics=[masked_perplexity])
 
     # 학습
     history = model.fit(dist_dataset, epochs=1, verbose=1)