diff --git a/research/cv/ICNet/src/models/icnet.py b/research/cv/ICNet/src/models/icnet.py
index 9973816cac0784cf1d0712fa1805105d9f0c6839..5b589b1eb09944b562949fbe3a8be85327e5b8e9 100644
--- a/research/cv/ICNet/src/models/icnet.py
+++ b/research/cv/ICNet/src/models/icnet.py
@@ -24,7 +24,7 @@ __all__ = ['ICNet']
 class ICNet(nn.Cell):
     """Image Cascade Network"""
 
-    def __init__(self, nclass=19, backbone='resnet50', pretrained_path='', istraining=True):
+    def __init__(self, nclass=19, backbone='resnet50', pretrained_path='', istraining=False):
         super(ICNet, self).__init__()
         self.conv_sub1 = nn.SequentialCell(
             _ConvBNReLU(3, 32, 3, 2),
@@ -34,7 +34,7 @@ class ICNet(nn.Cell):
         self.istraining = istraining
         self.ppm = PyramidPoolingModule()
 
-        self.backbone = SegBaseModel(root=pretrained_path)
+        self.backbone = SegBaseModel(root=pretrained_path, istraining=istraining)
 
         self.head = _ICHead(nclass)
 
@@ -228,11 +228,11 @@ class CascadeFeatureFusion24(nn.Cell):
 class SegBaseModel(nn.Cell):
     """Base Model for Semantic Segmentation"""
 
-    def __init__(self, nclass=19, backbone='resnet50', root=''):
+    def __init__(self, nclass=19, backbone='resnet50', root='', istraining=False):
         super(SegBaseModel, self).__init__()
         self.nclass = nclass
         if backbone == 'resnet50':
-            self.pretrained = get_resnet50v1b(ckpt_root=root)
+            self.pretrained = get_resnet50v1b(ckpt_root=root, istraining=istraining)
 
     def construct(self, x):
         """forwarding pre-trained network"""
diff --git a/research/cv/ICNet/src/models/icnet_dc.py b/research/cv/ICNet/src/models/icnet_dc.py
index 0ecd90ec82f7b7fbd92191a742648fb368a22ade..57d7bcc2136e5d6f442907f63cdeb4ae3cf6c9f2 100644
--- a/research/cv/ICNet/src/models/icnet_dc.py
+++ b/research/cv/ICNet/src/models/icnet_dc.py
@@ -34,7 +34,7 @@ class ICNetdc(nn.Cell):
         self.istraining = istraining
         self.ppm = PyramidPoolingModule()
 
-        self.backbone = SegBaseModel(root=pretrained_path)
+        self.backbone = SegBaseModel(root=pretrained_path, istraining=istraining)
 
         self.head = _ICHead(nclass, norm_layer=norm_layer)
 
@@ -230,11 +230,11 @@ class CascadeFeatureFusion24(nn.Cell):
 class SegBaseModel(nn.Cell):
     """Base Model for Semantic Segmentation"""
 
-    def __init__(self, nclass=19, backbone='resnet50', root=""):
+    def __init__(self, nclass=19, backbone='resnet50', root="", istraining=False):
         super(SegBaseModel, self).__init__()
         self.nclass = nclass
         if backbone == 'resnet50':
-            self.pretrained = get_resnet50v1b(ckpt_root=root)
+            self.pretrained = get_resnet50v1b(ckpt_root=root, istraining=istraining)
 
     def construct(self, x):
         """forwarding pre-trained network"""
diff --git a/research/cv/ICNet/src/models/resnet50_v1.py b/research/cv/ICNet/src/models/resnet50_v1.py
index b2a934fe75475f1b77cd1ad771777360b01b9153..c091ec5fa5361bea3f33ccbebd42f3c7d7bb9553 100644
--- a/research/cv/ICNet/src/models/resnet50_v1.py
+++ b/research/cv/ICNet/src/models/resnet50_v1.py
@@ -258,7 +258,7 @@ class Resnet50v1b(nn.Cell):
         return out
 
 
-def get_resnet50v1b(class_num=1001, ckpt_root='', pretrained=True):
+def get_resnet50v1b(class_num=1001, ckpt_root='', pretrained=True, istraining=False):
     """
     Get SE-ResNet50 neural network.
     Default : GE Theta+ version (best)
@@ -278,7 +278,7 @@ def get_resnet50v1b(class_num=1001, ckpt_root='', pretrained=True):
                         out_channels=[256, 512, 1024, 2048],
                         strides=[1, 2, 2, 2],
                         num_classes=class_num)
-
+    model.set_train(istraining)
     if pretrained and model.training:
         pretrained_ckpt = ckpt_root
         param_dict = load_checkpoint(pretrained_ckpt)