code cleanup for TF 0.12.1

ruimina · Feb 13, 2017 · b735da4 · b735da4
1 parent ed69cf7
commit b735da4
Show file tree

Hide file tree

Showing 3 changed files with 14 additions and 12 deletions.
diff --git a/train.py b/train.py
@@ -101,7 +101,8 @@ def _str_to_bool(s):
     parser.add_argument('--gc_channels', type=int, default=None,
                         help='Number of global condition channels. Default: None. Expecting: Int')
     parser.add_argument('--max_checkpoints', type=int, default=MAX_TO_KEEP,
-                        help='Maximum amount of checkpoints that will be kept alive. Default: ' + str(MAX_TO_KEEP) + '.')
+                        help='Maximum amount of checkpoints that will be kept alive. Default: '
+                             + str(MAX_TO_KEEP) + '.')
     return parser.parse_args()
 
 
@@ -195,7 +196,6 @@ def main():
         return
 
     logdir = directories['logdir']
-    logdir_root = directories['logdir_root']
     restore_from = directories['restore_from']
 
     # Even if we restored the model, we will treat it as new training
@@ -225,7 +225,7 @@ def main():
                                                                    wavenet_params["scalar_input"],
                                                                    wavenet_params["initial_filter_width"]),
             sample_size=args.sample_size,
-            silence_threshold=args.silence_threshold)
+            silence_threshold=silence_threshold)
         audio_batch = reader.dequeue(args.batch_size)
         if gc_enabled:
             gc_id_batch = reader.dequeue_gc(args.batch_size)
@@ -260,14 +260,14 @@ def main():
     optim = optimizer.minimize(loss, var_list=trainable)
 
     # Set up logging for TensorBoard.
-    writer = tf.train.SummaryWriter(logdir)
+    writer = tf.summary.FileWriter(logdir)
     writer.add_graph(tf.get_default_graph())
     run_metadata = tf.RunMetadata()
-    summaries = tf.merge_all_summaries()
+    summaries = tf.summary.merge_all()
 
     # Set up session
     sess = tf.Session(config=tf.ConfigProto(log_device_placement=False))
-    init = tf.initialize_all_variables()
+    init = tf.global_variables_initializer()
     sess.run(init)
 
     # Saver for storing checkpoints of the model.
@@ -290,8 +290,8 @@ def main():
     reader.start_threads(sess)
 
     step = None
+    last_saved_step = saved_global_step
     try:
-        last_saved_step = saved_global_step
         for step in range(saved_global_step + 1, args.num_steps):
             start_time = time.time()
             if args.store_metadata and step % 50 == 0:

diff --git a/wavenet/model.py b/wavenet/model.py
@@ -667,7 +667,7 @@ def loss(self,
                     target_output)
                 reduced_loss = tf.reduce_mean(loss)
 
-                tf.scalar_summary('loss', reduced_loss)
+                tf.summary.scalar('loss', reduced_loss)
 
                 if l2_regularization_strength is None:
                     return reduced_loss
@@ -681,7 +681,7 @@ def loss(self,
                     total_loss = (reduced_loss +
                                   l2_regularization_strength * l2_loss)
 
-                    tf.scalar_summary('l2_loss', l2_loss)
-                    tf.scalar_summary('total_loss', total_loss)
+                    tf.summary.scalar('l2_loss', l2_loss)
+                    tf.summary.scalar('total_loss', total_loss)
 
                     return total_loss
diff --git a/wavenet/ops.py b/wavenet/ops.py
@@ -66,8 +66,10 @@ def mu_law_encode(audio, quantization_channels):
     with tf.name_scope('encode'):
         mu = quantization_channels - 1
         # Perform mu-law companding transformation (ITU-T, 1988).
-        # Minimum operation is here to deal with rare large amplitudes caused by resampling.
-        magnitude = tf.log(1 + mu * tf.minimum(tf.abs(audio), 1.0)) / tf.log(1. + mu)
+        # Minimum operation is here to deal with rare large amplitudes caused
+        # by resampling.
+        safe_audio_abs = tf.minimum(tf.abs(audio), 1.0)
+        magnitude = tf.log(1 + mu * safe_audio_abs) / tf.log(1. + mu)
         signal = tf.sign(audio) * magnitude
         # Quantize signal to the specified number of levels.
         return tf.cast((signal + 1) / 2 * mu + 0.5, tf.int32)