sessinit paraemter

example_cifar10.py

@@ -13,9 +13,7 @@ from tensorpack.utils import *
 from tensorpack.utils.symbolic_functions import *
 from tensorpack.utils.summary import *
 from tensorpack.utils.callback import *
-from tensorpack.utils.sessinit import *
 from tensorpack.utils.validation_callback import *
-from tensorpack.dataflow.dataset import Cifar10
 from tensorpack.dataflow import *
 
 BATCH_SIZE = 128
@@ -81,10 +79,10 @@ def get_config():
     logger.set_logger_dir(log_dir)
 
     import cv2
-    dataset_train = Cifar10('train')
+    dataset_train = dataset.Cifar10('train')
     dataset_train = MapData(dataset_train, lambda img: cv2.resize(img, (24, 24)))
     dataset_train = BatchData(dataset_train, 128)
-    dataset_test = Cifar10('test')
+    dataset_test = dataset.Cifar10('test')
     dataset_test = MapData(dataset_test, lambda img: cv2.resize(img, (24, 24)))
     dataset_test = BatchData(dataset_test, 128)
     step_per_epoch = dataset_train.size()
@@ -133,7 +131,6 @@ if __name__ == '__main__':
     parser = argparse.ArgumentParser()
     parser.add_argument('--gpu', help='comma separated list of GPU(s) to use.') # nargs='*' in multi mode
     parser.add_argument('--load', help='load model')
-    global args
     args = parser.parse_args()
     if args.gpu:
         os.environ['CUDA_VISIBLE_DEVICES'] = args.gpu
@@ -141,8 +138,7 @@ if __name__ == '__main__':
     with tf.Graph().as_default():
         train.prepare()
         config = get_config()
-
         if args.load:
             config['session_init'] = SaverRestore(args.load)
-        sess_init = NewSession()
+
         train.start_train(config)

example_mnist.py

@@ -16,7 +16,6 @@ from tensorpack.utils.symbolic_functions import *
 from tensorpack.utils.summary import *
 from tensorpack.utils.callback import *
 from tensorpack.utils.validation_callback import *
-from tensorpack.dataflow.dataset import Mnist
 from tensorpack.dataflow import *
 
 BATCH_SIZE = 128
@@ -95,8 +94,8 @@ def get_config():
 
     IMAGE_SIZE = 28
 
-    dataset_train = BatchData(Mnist('train'), 128)
-    dataset_test = BatchData(Mnist('test'), 256, remainder=True)
+    dataset_train = BatchData(dataset.Mnist('train'), 128)
+    dataset_test = BatchData(dataset.Mnist('test'), 256, remainder=True)
     step_per_epoch = dataset_train.size()
     #step_per_epoch = 20
     #dataset_test = FixedSizeData(dataset_test, 20)
@@ -144,6 +143,7 @@ if __name__ == '__main__':
     from tensorpack import train
     parser = argparse.ArgumentParser()
     parser.add_argument('--gpu', help='comma separated list of GPU(s) to use.') # nargs='*' in multi mode
+    parser.add_argument('--load', help='load model')
     args = parser.parse_args()
     if args.gpu:
         os.environ['CUDA_VISIBLE_DEVICES'] = args.gpu
@@ -151,4 +151,6 @@ if __name__ == '__main__':
     with tf.Graph().as_default():
         train.prepare()
         config = get_config()
+        if args.load:
+            config['session_init'] = SaverRestore(args.load)
         train.start_train(config)

tensorpack/dataflow/__init__.py

@@ -6,6 +6,7 @@
 from pkgutil import walk_packages
 import os
 import os.path
+import dataset
 
 __SKIP = ['dftools', 'dataset']
 def global_import(name):

tensorpack/models/_common.py

@@ -4,6 +4,8 @@
 # Author: Yuxin Wu <ppwwyyxx@gmail.com>
 
 import tensorflow as tf
+from functools import wraps
+
 from ..utils.modelutils import *
 from ..utils.summary import *
 from ..utils import logger
@@ -21,6 +23,7 @@ def layer_register(summary_activation=False):
             Can be overriden when creating the layer.
     """
     def wrapper(func):
+        @wraps(func)
         def inner(*args, **kwargs):
             name = args[0]
             assert isinstance(name, basestring)

tensorpack/models/regularize.py

@@ -7,9 +7,14 @@ import tensorflow as tf
 import re
 
 from ..utils import logger
+from ..utils import *
 
 __all__ = ['regularize_cost']
 
+@memoized
+def _log_regularizer(name):
+    logger.info("Apply regularizer for {}".format(name))
+
 def regularize_cost(regex, func):
     """
     Apply a regularizer on every trainable variable matching the regex
@@ -20,8 +25,7 @@ def regularize_cost(regex, func):
     costs = []
     for p in params:
         name = p.name
-        if re.search(regex, name):
-            logger.info("Apply regularizer for {}".format(name))
-            costs.append(func(p))
+        costs.append(func(p))
+        _log_regularizer(name)
     return tf.add_n(costs)
 

tensorpack/train.py

@@ -11,7 +11,6 @@ from utils import *
 from utils.concurrency import EnqueueThread,coordinator_guard
 from utils.summary import summary_moving_average
 from utils.modelutils import describe_model
-from utils.sessinit import NewSession
 from utils import logger
 from dataflow import DataFlow
 

tensorpack/utils/__init__.py

@@ -8,8 +8,11 @@ import os
 import time
 import sys
 from contextlib import contextmanager
-import logger
 import tensorflow as tf
+import collections
+
+
+import logger
 
 def global_import(name):
     p = __import__(name, globals(), locals())
@@ -17,6 +20,7 @@ def global_import(name):
     for k in lst:
         globals()[k] = p.__dict__[k]
 global_import('naming')
+global_import('sessinit')
 
 @contextmanager
 def timed_operation(msg, log_start=False):
@@ -66,3 +70,31 @@ def get_default_sess_config():
     conf.allow_soft_placement = True
     return conf
 
+class memoized(object):
+    '''Decorator. Caches a function's return value each time it is called.
+    If called later with the same arguments, the cached value is returned
+    (not reevaluated).
+    '''
+    def __init__(self, func):
+       self.func = func
+       self.cache = {}
+
+    def __call__(self, *args):
+       if not isinstance(args, collections.Hashable):
+          # uncacheable. a list, for instance.
+          # better to not cache than blow up.
+          return self.func(*args)
+       if args in self.cache:
+          return self.cache[args]
+       else:
+          value = self.func(*args)
+          self.cache[args] = value
+          return value
+
+    def __repr__(self):
+       '''Return the function's docstring.'''
+       return self.func.__doc__
+
+    def __get__(self, obj, objtype):
+       '''Support instance methods.'''
+       return functools.partial(self.__call__, obj)

tensorpack/utils/sessinit.py

@@ -7,7 +7,7 @@ from abc import abstractmethod
 import numpy as np
 import tensorflow as tf
 
-from . import logger
+import logger
 class SessionInit(object):
     @abstractmethod
     def init(self, sess):

tensorpack/utils/summary.py

@@ -5,7 +5,7 @@
 
 import tensorflow as tf
 import logger
-from .naming import *
+from naming import *
 
 def create_summary(name, v):
     """