!3241 modify FaceRecognition and maskrcnn network scripts

Merge pull request !3241 from anzhengqi/modify-networks

official/cv/maskrcnn/src/maskrcnn/fpn_neck.py

@@ -55,7 +55,7 @@ class FeatPyramidNeck(nn.Cell):
         Tuple, with tensors of same channel size.
 
     Examples:
-        neck = FeatPyramidNeck([100,200,300], 50, 4)
+        neck = FeatPyramidNeck([100,200,300], 50, 4, config.feature_shapes)
         input_data = (normal(0,0.1,(1,c,1280//(4*2**i), 768//(4*2**i)),
                       dtype=np.float32) \
                       for i, c in enumerate(config.fpn_in_channels))
@@ -65,7 +65,8 @@ class FeatPyramidNeck(nn.Cell):
     def __init__(self,
                  in_channels,
                  out_channels,
-                 num_outs):
+                 num_outs,
+                 feature_shapes):
         super(FeatPyramidNeck, self).__init__()
 
         if context.get_context("device_target") == "Ascend":
@@ -91,9 +92,9 @@ class FeatPyramidNeck(nn.Cell):
             self.fpn_convs_.append(fpn_conv)
         self.lateral_convs_list = nn.layer.CellList(self.lateral_convs_list_)
         self.fpn_convs_list = nn.layer.CellList(self.fpn_convs_)
-        self.interpolate1 = P.ResizeBilinear((48, 80))
-        self.interpolate2 = P.ResizeBilinear((96, 160))
-        self.interpolate3 = P.ResizeBilinear((192, 320))
+        self.interpolate1 = P.ResizeBilinear(feature_shapes[2])
+        self.interpolate2 = P.ResizeBilinear(feature_shapes[1])
+        self.interpolate3 = P.ResizeBilinear(feature_shapes[0])
         self.cast = P.Cast()
         self.maxpool = P.MaxPool(kernel_size=1, strides=2, pad_mode="same")
 

official/cv/maskrcnn/src/maskrcnn/mask_rcnn_r50.py

@@ -96,7 +96,8 @@ class Mask_Rcnn_Resnet50(nn.Cell):
         # Fpn
         self.fpn_ncek = FeatPyramidNeck(config.fpn_in_channels,
                                         config.fpn_out_channels,
-                                        config.fpn_num_outs)
+                                        config.fpn_num_outs,
+                                        config.feature_shapes)
 
         # Rpn and rpn loss
         self.gt_labels_stage1 = Tensor(np.ones((self.train_batch_size, config.num_gts)).astype(np.uint8))

official/cv/maskrcnn/src/maskrcnn/proposal_generator.py

-# Copyright 2020-2021 Huawei Technologies Co., Ltd
+# Copyright 2020-2022 Huawei Technologies Co., Ltd
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -64,15 +64,14 @@ class Proposal(nn.Cell):
         self.target_means = target_means
         self.target_stds = target_stds
         self.use_sigmoid_cls = use_sigmoid_cls
+        self.reshape_shape = (-1, 1)
 
         if self.use_sigmoid_cls:
             self.cls_out_channels = num_classes - 1
             self.activation = P.Sigmoid()
-            self.reshape_shape = (-1, 1)
         else:
             self.cls_out_channels = num_classes
             self.activation = P.Softmax(axis=1)
-            self.reshape_shape = (-1, 2)
 
         if self.cls_out_channels <= 0:
             raise ValueError('num_classes={} is too small'.format(num_classes))

research/cv/FaceRecognition/base_config.yaml

@@ -27,7 +27,6 @@ backbone: "r100"
 use_se: 1
 emb_size: 512
 act_type: "relu"
-fp16: 1
 pre_bn: 1
 inference: 0
 use_drop: 1

research/cv/FaceRecognition/base_config_cpu.yaml

@@ -27,7 +27,6 @@ backbone: "r100"
 use_se: 1
 emb_size: 512
 act_type: "relu"
-fp16: 1
 pre_bn: 1
 inference: 0
 use_drop: 1

research/cv/FaceRecognition/beta_config.yaml

@@ -27,7 +27,6 @@ backbone: "r100"
 use_se: 0
 emb_size: 256
 act_type: "relu"
-fp16: 1
 pre_bn: 0
 inference: 0
 use_drop: 1

research/cv/FaceRecognition/beta_config_cpu.yaml

@@ -27,7 +27,6 @@ backbone: "r100"
 use_se: 0
 emb_size: 256
 act_type: "relu"
-fp16: 1
 pre_bn: 0
 inference: 0
 use_drop: 1

research/cv/FaceRecognition/default_config.yaml

@@ -27,7 +27,6 @@ backbone: "r100"
 use_se: 1
 emb_size: 512
 act_type: "relu"
-fp16: 1
 pre_bn: 1
 inference: 0
 use_drop: 1

research/cv/FaceRecognition/inference_config.yaml

@@ -29,7 +29,6 @@ backbone: "r100"
 use_se: 0
 emb_size: 256
 act_type: "relu"
-fp16: 1
 pre_bn: 0
 inference: 1
 use_drop: 0

research/cv/FaceRecognition/inference_config_cpu.yaml

@@ -29,7 +29,6 @@ backbone: "r100"
 use_se: 0
 emb_size: 256
 act_type: "relu"
-fp16: 1
 pre_bn: 0
 inference: 1
 use_drop: 0

research/cv/FaceRecognition/train.py

@@ -260,12 +260,11 @@ def run_train():
     network_1 = DistributedHelper(_backbone, margin_fc_1)
     config.logger.info('DistributedHelper----out----')
     config.logger.info('network fp16----in----')
-    if config.fp16 == 1:
     network_1.add_flags_recursive(fp16=True)
     config.logger.info('network fp16----out----')
 
     criterion_1 = get_loss(config)
-    if config.fp16 == 1 and config.model_parallel == 0:
+    if config.model_parallel == 0:
         criterion_1.add_flags_recursive(fp32=True)
 
     network_1 = load_pretrain(config, network_1)

research/cv/centernet/src/decode.py

@@ -148,7 +148,6 @@ class GatherFeatureByInd(nn.Cell):
         self.reshape = ops.Reshape()
         self.enable_cpu_gatherd = enable_cpu_gatherd
         if self.enable_cpu_gatherd:
-            self.value = Tensor(2, mstype.int32)
             self.gather_nd = ops.GatherD()
             self.expand_dims = ops.ExpandDims()
         else:
@@ -165,7 +164,7 @@ class GatherFeatureByInd(nn.Cell):
             # (b, J, K, N)
             index = self.expand_dims(ind, -1)
             index = self.tile(index, (1, 1, 1, N))
-            feat = self.gather_nd(feat, self.value, index)
+            feat = self.gather_nd(feat, 2, index)
         else:
             ind = self.reshape(ind, (-1, 1))
             ind_b = nn.Range(0, b * J, 1)()

research/cv/centernet/src/utils.py

@@ -134,7 +134,6 @@ class GatherFeature(nn.Cell):
         self.reshape = ops.Reshape()
         self.enable_cpu_gather = enable_cpu_gather
         if self.enable_cpu_gather:
-            self.value = Tensor(1, mstype.int32)
             self.gather_nd = ops.GatherD()
             self.expand_dims = ops.ExpandDims()
         else:
@@ -147,7 +146,7 @@ class GatherFeature(nn.Cell):
             # (b, N, c)
             index = self.expand_dims(ind, -1)
             index = self.tile(index, (1, 1, c))
-            feat = self.gather_nd(feat, self.value, index)
+            feat = self.gather_nd(feat, 1, index)
         else:
             # (b, N)->(b*N, 1)
             b, N = self.shape(ind)