deeplabv3 hwc & ssd off_optimize & resnet18

modified:   model_zoo/official/cv/deeplabv3/eval.py

model_zoo/official/cv/deeplabv3/eval.py

@@ -21,6 +21,7 @@ import cv2
 from mindspore import Tensor
 import mindspore.common.dtype as mstype
 import mindspore.nn as nn
+import mindspore.ops as ops
 from mindspore import context
 from mindspore.train.serialization import load_checkpoint, load_param_into_net
 from src.nets import net_factory
@@ -47,6 +48,8 @@ def parse_args():
     parser.add_argument('--model', type=str, default='deeplab_v3_s16', help='select model')
     parser.add_argument('--freeze_bn', action='store_true', default=False, help='freeze bn')
     parser.add_argument('--ckpt_path', type=str, default='', help='model to evaluate')
+    parser.add_argument("--input_format", type=str, choices=["NCHW", "NHWC"], default="NCHW",
+                        help="NCHW or NHWC")
 
     args, _ = parser.parse_known_args()
     return args
@@ -70,12 +73,16 @@ def resize_long(img, long_size=513):
 
 
 class BuildEvalNetwork(nn.Cell):
-    def __init__(self, network):
+    def __init__(self, network, input_format="NCHW"):
         super(BuildEvalNetwork, self).__init__()
         self.network = network
         self.softmax = nn.Softmax(axis=1)
+        self.transpose = ops.Transpose()
+        self.format = input_format
 
     def construct(self, input_data):
+        if self.format == "NHWC":
+            input_data = self.transpose(input_data, (0, 3, 1, 2))
         output = self.network(input_data)
         output = self.softmax(output)
         return output
@@ -96,7 +103,6 @@ def pre_process(args, img_, crop_size=513):
     pad_w = crop_size - img_.shape[1]
     if pad_h > 0 or pad_w > 0:
         img_ = cv2.copyMakeBorder(img_, 0, pad_h, 0, pad_w, cv2.BORDER_CONSTANT, value=0)
-
     # hwc to chw
     img_ = img_.transpose((2, 0, 1))
     return img_, resize_h, resize_w
@@ -162,7 +168,7 @@ def net_eval():
     else:
         raise NotImplementedError('model [{:s}] not recognized'.format(args.model))
 
-    eval_net = BuildEvalNetwork(network)
+    eval_net = BuildEvalNetwork(network, args.input_format)
 
     # load model
     param_dict = load_checkpoint(args.ckpt_path)

model_zoo/official/cv/deeplabv3/export.py

@@ -32,6 +32,8 @@ parser.add_argument("--device_target", type=str, choices=["Ascend", "GPU", "CPU"
 parser.add_argument('--model', type=str.lower, default='deeplab_v3_s8', choices=['deeplab_v3_s16', 'deeplab_v3_s8'],
                     help='Select model structure (Default: deeplab_v3_s8)')
 parser.add_argument('--num_classes', type=int, default=21, help='the number of classes (Default: 21)')
+parser.add_argument("--input_format", type=str, choices=["NCHW", "NHWC"], default="NCHW",
+                    help="NCHW or NHWC")
 args = parser.parse_args()
 
 context.set_context(mode=context.GRAPH_MODE, device_target=args.device_target)
@@ -43,10 +45,13 @@ if __name__ == '__main__':
         network = net_factory.nets_map['deeplab_v3_s16']('eval', args.num_classes, 16, True)
     else:
         network = net_factory.nets_map['deeplab_v3_s8']('eval', args.num_classes, 8, True)
-    network = BuildEvalNetwork(network)
+    network = BuildEvalNetwork(network, args.input_format)
     param_dict = load_checkpoint(args.ckpt_file)
 
     # load the parameter into net
     load_param_into_net(network, param_dict)
-    input_data = Tensor(np.ones([args.batch_size, 3, args.input_size, args.input_size]).astype(np.float32))
+    if args.input_format == "NHWC":
+        input_data = Tensor(np.ones([args.batch_size, args.input_size, args.input_size, 3]).astype(np.float32))
+    else:
+        input_data = Tensor(np.ones([args.batch_size, 3, args.input_size, args.input_size]).astype(np.float32))
     export(network, input_data, file_name=args.file_name, file_format=args.file_format)

model_zoo/official/cv/resnet/ascend310_infer/src/main.cc

@@ -39,10 +39,9 @@ using mindspore::dataset::vision::CenterCrop;
 using mindspore::dataset::vision::Normalize;
 using mindspore::dataset::vision::HWC2CHW;
 using mindspore::dataset::TensorTransform;
-using mindspore::GlobalContext;
+using mindspore::Context;
 using mindspore::Serialization;
 using mindspore::Model;
-using mindspore::ModelContext;
 using mindspore::Status;
 using mindspore::ModelType;
 using mindspore::GraphCell;
@@ -62,14 +61,15 @@ int main(int argc, char **argv) {
     return 1;
   }
 
+  auto context = std::make_shared<Context>();
+  auto ascend310 = std::make_shared<mindspore::Ascend310DeviceInfo>();
+  ascend310->SetDeviceID(FLAGS_device_id);
+  context->MutableDeviceInfo().push_back(ascend310);
+  mindspore::Graph graph;
+  Serialization::Load(FLAGS_mindir_path, ModelType::kMindIR, &graph);
+  Model model;
+  Status ret = model.Build(GraphCell(graph), context);
 
-  GlobalContext::SetGlobalDeviceTarget(mindspore::kDeviceTypeAscend310);
-  GlobalContext::SetGlobalDeviceID(FLAGS_device_id);
-  auto graph = Serialization::LoadModel(FLAGS_mindir_path, ModelType::kMindIR);
-  auto model_context = std::make_shared<mindspore::ModelContext>();
-
-  Model model(GraphCell(graph), model_context);
-  Status ret = model.Build();
   if (ret != kSuccess) {
     std::cout << "ERROR: Build failed." << std::endl;
     return 1;

model_zoo/official/cv/ssd/ascend310_infer/src/main.cc

@@ -66,6 +66,7 @@ int main(int argc, char **argv) {
   auto context = std::make_shared<Context>();
   auto ascend310 = std::make_shared<mindspore::Ascend310DeviceInfo>();
   ascend310->SetDeviceID(FLAGS_device_id);
+  ascend310->SetBufferOptimizeMode("off_optimize");
   context->MutableDeviceInfo().push_back(ascend310);
   mindspore::Graph graph;
   Serialization::Load(FLAGS_mindir_path, ModelType::kMindIR, &graph);