replace step_per_epoch by steps_per_epoch in examples

docs/casestudies/colorize.md

@@ -298,9 +298,9 @@ def get_config():
     return TrainConfig(
         dataflow=dataset,
         optimizer=tf.train.AdamOptimizer(lr),
-        callbacks=Callbacks([StatPrinter(), PeriodicCallback(ModelSaver(), 3)])]),
+        callbacks=[PeriodicCallback(ModelSaver(), 3)],
         model=Model(),
-        step_per_epoch=dataset.size(),
+        steps_per_epoch=dataset.size(),
         max_epoch=100,
     )
 ```

examples/A3C-Gym/train-atari.py

@@ -219,7 +219,7 @@ def get_config():
         ],
         session_config=get_default_sess_config(0.5),
         model=M,
-        step_per_epoch=STEP_PER_EPOCH,
+        steps_per_epoch=STEP_PER_EPOCH,
         max_epoch=1000,
     )
 

examples/CTC-TIMIT/train-timit.py

@@ -89,7 +89,7 @@ def get_data(path, isTrain, stat_file):
 
 
 def get_config(ds_train, ds_test):
-    step_per_epoch = ds_train.size()
+    steps_per_epoch = ds_train.size()
 
     lr = symbolic_functions.get_scalar_var('learning_rate', 5e-3, summary=True)
 
@@ -105,7 +105,7 @@ def get_config(ds_train, ds_test):
                 InferenceRunner(ds_test, [ScalarStats('error')]), 2),
         ],
         model=Model(),
-        step_per_epoch=step_per_epoch,
+        steps_per_epoch=steps_per_epoch,
         max_epoch=70,
     )
 

examples/Char-RNN/char-rnn.py

@@ -103,7 +103,7 @@ def get_config():
 
     ds = CharRNNData(param.corpus, 100000)
     ds = BatchData(ds, param.batch_size)
-    step_per_epoch = ds.size()
+    steps_per_epoch = ds.size()
 
     lr = symbolic_functions.get_scalar_var('learning_rate', 2e-3, summary=True)
 
@@ -115,7 +115,7 @@ def get_config():
             ScheduledHyperParamSetter('learning_rate', [(25, 2e-4)])
         ],
         model=Model(),
-        step_per_epoch=step_per_epoch,
+        steps_per_epoch=steps_per_epoch,
         max_epoch=50,
     )
 

examples/DeepQNetwork/DQN.py

@@ -190,7 +190,7 @@ def get_config():
         # save memory for multiprocess evaluator
         session_config=get_default_sess_config(0.6),
         model=M,
-        step_per_epoch=STEP_PER_EPOCH,
+        steps_per_epoch=STEP_PER_EPOCH,
     )
 
 

examples/DoReFa-Net/alexnet-dorefa.py

@@ -247,7 +247,7 @@ def get_config():
                              ClassificationError('wrong-top5', 'val-error-top5')])
         ],
         model=Model(),
-        step_per_epoch=10000,
+        steps_per_epoch=10000,
         max_epoch=100,
     )
 

examples/DoReFa-Net/svhn-digit-dorefa.py

@@ -147,7 +147,7 @@ def get_config():
     data_train = AugmentImageComponent(data_train, augmentors)
     data_train = BatchData(data_train, 128)
     data_train = PrefetchDataZMQ(data_train, 5)
-    step_per_epoch = data_train.size()
+    steps_per_epoch = data_train.size()
 
     augmentors = [imgaug.Resize((40, 40))]
     data_test = AugmentImageComponent(data_test, augmentors)
@@ -169,7 +169,7 @@ def get_config():
                             [ScalarStats('cost'), ClassificationError()])
         ],
         model=Model(),
-        step_per_epoch=step_per_epoch,
+        steps_per_epoch=steps_per_epoch,
         max_epoch=200,
     )
 

examples/GAN/DCGAN-CelebA.py

@@ -112,7 +112,7 @@ def get_config():
         callbacks=[ModelSaver()],
         session_config=get_default_sess_config(0.5),
         model=Model(),
-        step_per_epoch=300,
+        steps_per_epoch=300,
         max_epoch=200,
     )
 

examples/GAN/Image2Image.py

@@ -173,7 +173,7 @@ def get_config():
             ScheduledHyperParamSetter('learning_rate', [(200, 1e-4)])
         ],
         model=Model(),
-        step_per_epoch=dataset.size(),
+        steps_per_epoch=dataset.size(),
         max_epoch=300,
     )
 

examples/GAN/InfoGAN-mnist.py

@@ -161,7 +161,7 @@ def get_config():
         callbacks=[ModelSaver()],
         session_config=get_default_sess_config(0.5),
         model=Model(),
-        step_per_epoch=500,
+        steps_per_epoch=500,
         max_epoch=100,
     )
 

examples/HED/hed.py

@@ -166,7 +166,7 @@ def view_data():
 def get_config():
     logger.auto_set_dir()
     dataset_train = get_data('train')
-    step_per_epoch = dataset_train.size() * 40
+    steps_per_epoch = dataset_train.size() * 40
     dataset_val = get_data('val')
 
     lr = get_scalar_var('learning_rate', 3e-5, summary=True)
@@ -181,7 +181,7 @@ def get_config():
                             BinaryClassificationStats('prediction', 'edgemap4d'))
         ],
         model=Model(),
-        step_per_epoch=step_per_epoch,
+        steps_per_epoch=steps_per_epoch,
         max_epoch=100,
     )
 

examples/Inception/inception-bn.py

@@ -153,7 +153,7 @@ def get_config():
     logger.auto_set_dir()
     # prepare dataset
     dataset_train = get_data('train')
-    step_per_epoch = 5000
+    steps_per_epoch = 5000
     dataset_val = get_data('val')
 
     lr = get_scalar_var('learning_rate', 0.045, summary=True)
@@ -172,7 +172,7 @@ def get_config():
         ],
         session_config=get_default_sess_config(0.99),
         model=Model(),
-        step_per_epoch=step_per_epoch,
+        steps_per_epoch=steps_per_epoch,
         max_epoch=80,
     )
 

examples/Inception/inceptionv3.py

@@ -281,7 +281,7 @@ def get_config():
         ],
         session_config=get_default_sess_config(0.9),
         model=Model(),
-        step_per_epoch=5000,
+        steps_per_epoch=5000,
         max_epoch=100,
     )
 

examples/PennTreebank/PTB-LSTM.py

@@ -110,11 +110,11 @@ def get_config():
     data3, wd2id = get_PennTreeBank()
     global VOCAB_SIZE
     VOCAB_SIZE = len(wd2id)
-    step_per_epoch = (data3[0].shape[0] // BATCH - 1) // SEQ_LEN
+    steps_per_epoch = (data3[0].shape[0] // BATCH - 1) // SEQ_LEN
 
     train_data = TensorInput(
         lambda: ptb_producer(data3[0], BATCH, SEQ_LEN),
-        step_per_epoch)
+        steps_per_epoch)
     val_data = TensorInput(
         lambda: ptb_producer(data3[1], BATCH, SEQ_LEN),
         (data3[1].shape[0] // BATCH - 1) // SEQ_LEN)

examples/ResNet/cifar10-resnet.py

@@ -134,7 +134,7 @@ def get_config():
 
     # prepare dataset
     dataset_train = get_data('train')
-    step_per_epoch = dataset_train.size()
+    steps_per_epoch = dataset_train.size()
     dataset_test = get_data('test')
 
     lr = get_scalar_var('learning_rate', 0.01, summary=True)
@@ -149,7 +149,7 @@ def get_config():
                                       [(1, 0.1), (82, 0.01), (123, 0.001), (300, 0.0002)])
         ],
         model=Model(n=NUM_UNITS),
-        step_per_epoch=step_per_epoch,
+        steps_per_epoch=steps_per_epoch,
         max_epoch=400,
     )
 

examples/ResNet/imagenet-resnet.py

@@ -199,7 +199,7 @@ def get_config():
             HumanHyperParamSetter('learning_rate'),
         ],
         model=Model(),
-        step_per_epoch=5000,
+        steps_per_epoch=5000,
         max_epoch=110,
     )
 

examples/ResNet/svhn-resnet.py

@@ -63,7 +63,7 @@ def get_config():
 
     # prepare dataset
     dataset_train = get_data('train')
-    step_per_epoch = dataset_train.size()
+    steps_per_epoch = dataset_train.size()
     dataset_test = get_data('test')
 
     lr = get_scalar_var('learning_rate', 0.01, summary=True)
@@ -78,7 +78,7 @@ def get_config():
                                       [(1, 0.1), (20, 0.01), (28, 0.001), (50, 0.0001)])
         ],
         model=Model(n=18),
-        step_per_epoch=step_per_epoch,
+        steps_per_epoch=steps_per_epoch,
         max_epoch=500,
     )
 

examples/SimilarityLearning/mnist-embeddings.py

@@ -133,7 +133,7 @@ def get_config(model):
     logger.auto_set_dir()
 
     dataset = model.get_data()
-    step_per_epoch = dataset.size()
+    steps_per_epoch = dataset.size()
 
     lr = symbf.get_scalar_var('learning_rate', 1e-4, summary=True)
 
@@ -145,7 +145,7 @@ def get_config(model):
             ModelSaver(),
             ScheduledHyperParamSetter('learning_rate', [(10, 1e-5), (20, 1e-6)])
         ],
-        step_per_epoch=step_per_epoch,
+        steps_per_epoch=steps_per_epoch,
         max_epoch=20,
     )
 

examples/SpatialTransformer/mnist-addition.py

@@ -148,7 +148,7 @@ def get_config():
     logger.auto_set_dir()
 
     dataset_train, dataset_test = get_data(True), get_data(False)
-    step_per_epoch = dataset_train.size() * 5
+    steps_per_epoch = dataset_train.size() * 5
 
     lr = symbf.get_scalar_var('learning_rate', 5e-4, summary=True)
 
@@ -163,7 +163,7 @@ def get_config():
         ],
         session_config=get_default_sess_config(0.5),
         model=Model(),
-        step_per_epoch=step_per_epoch,
+        steps_per_epoch=steps_per_epoch,
         max_epoch=500,
     )
 

examples/cifar-convnet.py

@@ -107,7 +107,7 @@ def get_config(cifar_classnum):
 
     # prepare dataset
     dataset_train = get_data('train', cifar_classnum)
-    step_per_epoch = dataset_train.size()
+    steps_per_epoch = dataset_train.size()
     dataset_test = get_data('test', cifar_classnum)
 
     sess_config = get_default_sess_config(0.5)
@@ -130,7 +130,7 @@ def get_config(cifar_classnum):
         ],
         session_config=sess_config,
         model=Model(cifar_classnum),
-        step_per_epoch=step_per_epoch,
+        steps_per_epoch=steps_per_epoch,
         max_epoch=150,
     )
 

examples/mnist-convnet.py

@@ -124,7 +124,7 @@ def get_config():
 
     dataset_train, dataset_test = get_data()
     # how many iterations you want in each epoch
-    step_per_epoch = dataset_train.size()
+    steps_per_epoch = dataset_train.size()
 
     lr = tf.train.exponential_decay(
         learning_rate=1e-3,
@@ -148,7 +148,7 @@ def get_config():
                 [ScalarStats('cross_entropy_loss'), ClassificationError('incorrect')]),
         ],
         model=Model(),
-        step_per_epoch=step_per_epoch,
+        steps_per_epoch=steps_per_epoch,
         max_epoch=100,
     )
 

examples/svhn-digit-convnet.py

@@ -89,7 +89,7 @@ def get_config():
     logger.auto_set_dir()
 
     data_train, data_test = get_data()
-    step_per_epoch = data_train.size()
+    steps_per_epoch = data_train.size()
 
     lr = tf.train.exponential_decay(
         learning_rate=1e-3,
@@ -107,7 +107,7 @@ def get_config():
                             [ScalarStats('cost'), ClassificationError()])
         ],
         model=Model(),
-        step_per_epoch=step_per_epoch,
+        steps_per_epoch=steps_per_epoch,
         max_epoch=350,
     )
 

tensorpack/models/model_desc.py

@@ -5,7 +5,6 @@
 
 from abc import ABCMeta, abstractmethod
 import tensorflow as tf
-import inspect
 import pickle
 import six
 
@@ -102,12 +101,7 @@ class ModelDesc(object):
             model_inputs (list[tf.Tensor]): a list of inputs, corresponding to
                 InputVars of this model.
         """
-        if len(inspect.getargspec(self._build_graph).args) == 3:
-            logger.warn("[DEPRECATED] _build_graph(self, input_vars, is_training) is deprecated! \
-Use _build_graph(self, input_vars) and get_current_tower_context().is_training instead.")
-            self._build_graph(model_inputs, get_current_tower_context().is_training)
-        else:
-            self._build_graph(model_inputs)
+        self._build_graph(model_inputs)
 
     @abstractmethod
     def _build_graph(self, inputs):