Fix some more scope issues

examples/FasterRCNN/basemodel.py

@@ -6,6 +6,7 @@ from contextlib import contextmanager
 import tensorflow as tf
 from tensorpack.tfutils.argscope import argscope, get_arg_scope
 from tensorpack.tfutils.scope_utils import auto_reuse_variable_scope
+from tensorpack.tfutils.varreplace import custom_getter_scope
 from tensorpack.models import (
     Conv2D, MaxPooling, BatchNorm, BNReLU)
 
@@ -26,8 +27,7 @@ def resnet_argscope():
     with argscope([Conv2D, MaxPooling, BatchNorm], data_format='NCHW'), \
             argscope(Conv2D, use_bias=False), \
             argscope(BatchNorm, use_local_stat=False), \
-            tf.variable_scope(tf.get_variable_scope(),
-                              custom_getter=maybe_freeze_affine):
+            custom_getter_scope(maybe_freeze_affine):
         yield
 
 

examples/FasterRCNN/model.py

@@ -503,7 +503,8 @@ def maskrcnn_head(feature, num_class):
     """
     with argscope([Conv2D, Deconv2D], data_format='NCHW',
                   W_init=tf.variance_scaling_initializer(
-                      scale=2.0, mode='fan_in', distribution='normal')):
+                      scale=2.0, mode='fan_out', distribution='normal')):
+        # c2's MSRAFill is fan_out
         l = Deconv2D('deconv', feature, 256, 2, stride=2, nl=tf.nn.relu)
         l = Conv2D('conv', l, num_class - 1, 1)
     return l

tensorpack/tfutils/scope_utils.py

@@ -8,6 +8,7 @@ import functools
 from contextlib import contextmanager
 
 from ..utils.argtools import graph_memoized
+from .common import get_tf_version_number
 
 __all__ = ['auto_reuse_variable_scope', 'cached_name_scope', 'under_name_scope']
 
@@ -39,8 +40,14 @@ def auto_reuse_variable_scope(func):
         h = hash((tf.get_default_graph(), scope.name))
         # print("Entering " + scope.name + " reuse: " + str(h in used_scope))
         if h in used_scope:
-            with tf.variable_scope(scope, reuse=True):
-                return func(*args, **kwargs)
+            if get_tf_version_number() >= 1.5:
+                with tf.variable_scope(scope, reuse=True, auxiliary_name_scope=False):
+                    return func(*args, **kwargs)
+            else:
+                ns = tf.get_default_graph().get_name_scope()
+                with tf.variable_scope(scope, reuse=True), \
+                        tf.name_scope(ns + '/' if ns else ''):
+                    return func(*args, **kwargs)
         else:
             used_scope.add(h)
             return func(*args, **kwargs)