better shape logging, add GAP

tensorpack/models/_common.py

@@ -39,12 +39,12 @@ def layer_register(summary_activation=False):
                 inputs = args[0]
                 with tf.variable_scope(name) as scope:
                     outputs = self.f(*args, **kwargs)
-                    if name not in _layer_logged:
+                    if scope.name not in _layer_logged:
                         # log shape info and add activation
                         logger.info("{} input: {}".format(
-                            name, get_shape_str(inputs)))
+                            scope.name, get_shape_str(inputs)))
                         logger.info("{} output: {}".format(
-                            name, get_shape_str(outputs)))
+                            scope.name, get_shape_str(outputs)))
 
                         if do_summary:
                             if isinstance(outputs, list):
@@ -52,7 +52,7 @@ def layer_register(summary_activation=False):
                                     add_activation_summary(x, scope.name)
                             else:
                                 add_activation_summary(outputs, scope.name)
-                        _layer_logged.add(name)
+                        _layer_logged.add(scope.name)
                     return outputs
         return WrapedObject(func)
     return wrapper

tensorpack/models/pool.py

@@ -8,7 +8,7 @@ import numpy
 from ._common import *
 from ..utils.symbolic_functions import *
 
-__all__ = ['MaxPooling', 'FixedUnPooling']
+__all__ = ['MaxPooling', 'FixedUnPooling', 'AvgPooling', 'GlobalAvgPooling']
 
 @layer_register()
 def MaxPooling(x, shape, stride=None, padding='VALID'):
@@ -26,6 +26,26 @@ def MaxPooling(x, shape, stride=None, padding='VALID'):
 
     return tf.nn.max_pool(x, ksize=shape, strides=stride, padding=padding)
 
+@layer_register()
+def AvgPooling(x, shape, stride=None, padding='VALID'):
+    """
+        shape, stride: int or list/tuple of length 2
+        if stride is None, use shape by default
+        padding: 'VALID' or 'SAME'
+    """
+    padding = padding.upper()
+    shape = shape4d(shape)
+    if stride is None:
+        stride = shape
+    else:
+        stride = shape4d(stride)
+
+    return tf.nn.avg_pool(x, ksize=shape, strides=stride, padding=padding)
+
+@layer_register()
+def GlobalAvgPooling(x):
+    assert x.get_shape().ndims == 4
+    return tf.reduce_mean(x, [1, 2])
 
 @layer_register()
 def FixedUnPooling(x, shape, unpool_mat=None):

tensorpack/utils/logger.py

@@ -41,13 +41,16 @@ def getlogger():
     logger.addHandler(handler)
     return logger
 
+def get_time_str():
+    return datetime.now().strftime('%m%d-%H%M%S')
+
 logger = getlogger()
 
 # logger file and directory:
 global LOG_FILE, LOG_DIR
 def _set_file(path):
     if os.path.isfile(path):
-        backup_name = path + datetime.now().strftime('.%d-%H%M%S')
+        backup_name = path + '.' + get_time_str()
         shutil.move(path, backup_name)
         info("Log file '{}' backuped to '{}'".format(path, backup_name))
     hdl = logging.FileHandler(
@@ -65,15 +68,14 @@ unless you're resuming from a previous task.""".format(dirname))
             act = input().lower()
             if act:
                 break
-        timestr = datetime.now().strftime('%m%d-%H%M%S')
         if act == 'b':
-            backup_name = dirname + timestr
+            backup_name = dirname + get_time_str()
             shutil.move(dirname, backup_name)
             info("Directory'{}' backuped to '{}'".format(dirname, backup_name))
         elif act == 'd':
             shutil.rmtree(dirname)
         elif act == 'n':
-            dirname = dirname + timestr
+            dirname = dirname + get_time_str()
             info("Use a different log directory {}".format(dirname))
         elif act == 'k':
             pass