Committed caffe-tensorflow modifications to PSPNet BN layer

caffe-tensorflow/kaffe/layers.py

@@ -11,6 +11,7 @@ LAYER_DESCRIPTORS = {
     'Accuracy': shape_scalar,
     'ArgMax': shape_not_implemented,
     'BatchNorm': shape_identity,
+    'BN': shape_identity,
     'BNLL': shape_not_implemented,
     'Concat': shape_concat,
     'ContrastiveLoss': shape_scalar,
@@ -18,12 +19,13 @@ LAYER_DESCRIPTORS = {
     'Deconvolution': shape_not_implemented,
     'Data': shape_data,
     'Dropout': shape_identity,
+    'Interp': shape_identity,
     'DummyData': shape_data,
     'EuclideanLoss': shape_scalar,
     'Eltwise': shape_identity,
     'Exp': shape_identity,
     'Flatten': shape_not_implemented,
-    'HDF5Data': shape_data,
+    'HDF5Datapython': shape_data,
     'HDF5Output': shape_identity,
     'HingeLoss': shape_scalar,
     'Im2col': shape_not_implemented,

caffe-tensorflow/kaffe/tensorflow/transformer.py

@@ -104,6 +104,9 @@ class TensorFlowMapper(NodeMapper):
     def map_relu(self, node):
         return TensorFlowNode('relu')
 
+    def map_interp(self, node):
+        return TensorFlowNode('relu')
+
     def map_pooling(self, node):
         pool_type = node.parameters.pool
         if pool_type == 0:
@@ -150,6 +153,12 @@ class TensorFlowMapper(NodeMapper):
         kwargs = {} if scale_offset else {'scale_offset': False}
         return MaybeActivated(node, default=False)('batch_normalization', **kwargs)
 
+    def map_bn(self, node):
+        scale_offset = len(node.data) == 4
+        print node
+        kwargs = {} if scale_offset else {'scale_offset': False}
+        return MaybeActivated(node, default=False)('batch_normalization', **kwargs)
+
     def map_eltwise(self, node):
         operations = {0: 'multiply', 1: 'add', 2: 'max'}
         op_code = node.parameters.operation

caffe-tensorflow/kaffe/transformers.py

@@ -119,6 +119,7 @@ class DataReshaper(object):
             if node.data is None:
                 continue
             if node.kind not in self.reshaped_node_types:
+
                 # Check for 2+ dimensional data
                 if any(len(tensor.shape) > 1 for tensor in node.data):
                     print_stderr('Warning: parmaters not reshaped for node: {}'.format(node))
@@ -155,6 +156,7 @@ class SubNodeFuser(object):
 
     def __call__(self, graph):
         nodes = graph.nodes
+        print graph.name
         fused_nodes = []
         for node in nodes:
             if len(node.parents) != 1:
@@ -216,11 +218,21 @@ class BatchNormScaleBiasFuser(SubNodeFuser):
     This fuser merges the scaling+bias layer with the batch norm.
     '''
 
-    def is_eligible_pair(self, parent, child):
+    def is_eligible_pair_(self, parent, child):
         return (parent.kind == NodeKind.BatchNorm and child.kind == NodeKind.Scale and
                 child.parameters.axis == 1 and child.parameters.bias_term == True)
 
+    '''
+    Made for the purpose of 
+    '''
+    def is_eligible_pair(self, parent, child):
+        if parent.kind == NodeKind.BatchNorm: 
+            print 'kaffe/transformers.py  line 227: use function above'
+        return (parent.kind == NodeKind.BN and child.kind == NodeKind.Scale and
+                child.parameters.axis == 1 and child.parameters.bias_term == True)
+
     def merge(self, parent, child):
+
         parent.scale_bias_node = child
 
 
@@ -232,10 +244,15 @@ class BatchNormPreprocessor(object):
 
     def __call__(self, graph):
         for node in graph.nodes:
-            if node.kind != NodeKind.BatchNorm:
+            if node.kind not in (NodeKind.BatchNorm, NodeKind.BN):
+                continue
+            else:
                 continue
             assert node.data is not None
+            print len(node.data)
+            print node
             assert len(node.data) == 3
+
             mean, variance, scale = node.data
             # Prescale the stats
             scaling_factor = 1.0 / scale if scale != 0 else 0
@@ -278,9 +295,10 @@ class ParameterNamer(object):
                 names = ('weights',)
                 if node.parameters.bias_term:
                     names += ('biases',)
-            elif node.kind == NodeKind.BatchNorm:
+            elif node.kind in (NodeKind.BatchNorm, NodeKind.BN):
                 names = ('mean', 'variance')
                 if len(node.data) == 4:
+
                     names += ('scale', 'offset')
             else:
                 print_stderr('WARNING: Unhandled parameters: {}'.format(node.kind))