Select backbone ResNet50/101 in config.

config.py

@@ -48,6 +48,10 @@ class Config(object):
     # down the training.
     VALIDATION_STEPS = 50
 
+    # Backbone network architecture
+    # Supported values are: resnet50, resnet101
+    BACKBONE = "resnet101"
+
     # The strides of each layer of the FPN Pyramid. These values
     # are based on a Resnet101 backbone.
     BACKBONE_STRIDES = [4, 8, 16, 32, 64]

model.py

@@ -1821,7 +1821,7 @@ class MaskRCNN():
         # Bottom-up Layers
         # Returns a list of the last layers of each stage, 5 in total.
         # Don't create the thead (stage 5), so we pick the 4th item in the list.
-        _, C2, C3, C4, C5 = resnet_graph(input_image, "resnet101", stage5=True)
+        _, C2, C3, C4, C5 = resnet_graph(input_image, config.BACKBONE, stage5=True)
         # Top-down Layers
         # TODO: add assert to varify feature map sizes match what's in config
         P5 = KL.Conv2D(256, (1, 1), name='fpn_c5p5')(C5)