diff --git a/model_zoo/official/cv/efficientnet/export.py b/model_zoo/official/cv/efficientnet/export.py
index 0c041a2900a..f53dc2b7e6d 100644
--- a/model_zoo/official/cv/efficientnet/export.py
+++ b/model_zoo/official/cv/efficientnet/export.py
@@ -35,7 +35,7 @@ args = parser.parse_args()
 context.set_context(mode=context.GRAPH_MODE, device_target=args.device_target, device_id=args.device_id)
 
 if __name__ == "__main__":
-    if args.platform != "GPU":
+    if args.device_target != "GPU":
         raise ValueError("Only supported GPU now.")
 
     net = efficientnet_b0(num_classes=cfg.num_classes,
diff --git a/model_zoo/official/gnn/gat/export.py b/model_zoo/official/gnn/gat/export.py
index 686580773be..0a7ce00bb90 100644
--- a/model_zoo/official/gnn/gat/export.py
+++ b/model_zoo/official/gnn/gat/export.py
@@ -12,26 +12,31 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 # ============================================================================
-"""export checkpoint file into air models"""
+"""export checkpoint file into air, mindir and onnx models"""
 import argparse
 import numpy as np
 
-from mindspore import Tensor, context
-from mindspore.train.serialization import load_checkpoint, export
+from mindspore import Tensor, context, load_checkpoint, export
 
 from src.gat import GAT
 from src.config import GatConfig
 
-context.set_context(mode=context.GRAPH_MODE, device_target="Ascend")
+parser = argparse.ArgumentParser(description="GAT export")
+parser.add_argument("--device_id", type=int, default=0, help="Device id")
+parser.add_argument("--ckpt_file", type=str, required=True, help="Checkpoint file path.")
+parser.add_argument("--dataset", type=str, default="cora", choices=["cora", "citeseer"], help="Dataset.")
+parser.add_argument("--file_name", type=str, default="gat", help="output file name.")
+parser.add_argument("--file_format", type=str, choices=["AIR", "ONNX", "MINDIR"], default="AIR", help="file format")
+parser.add_argument("--device_target", type=str, default="Ascend",
+                    choices=["Ascend", "GPU", "CPU"], help="device target (default: Ascend)")
+args = parser.parse_args()
 
-if __name__ == '__main__':
-    parser = argparse.ArgumentParser(description='GAT_export')
-    parser.add_argument('--ckpt_file', type=str, default='./ckpts/gat.ckpt', help='GAT ckpt file.')
-    parser.add_argument('--output_file', type=str, default='gat.air', help='GAT output air name.')
-    parser.add_argument('--dataset', type=str, default='cora', help='GAT dataset name.')
-    args_opt = parser.parse_args()
+context.set_context(mode=context.GRAPH_MODE, device_target=args.device_target, device_id=args.device_id)
 
-    if args_opt.dataset == "citeseer":
+if __name__ == "__main__":
+
+
+    if args.dataset == "citeseer":
         feature_size = [1, 3312, 3703]
         biases_size = [1, 3312, 3312]
         num_classes = 6
@@ -58,7 +63,7 @@ if __name__ == '__main__':
                   ftr_drop=0.0)
 
     gat_net.set_train(False)
-    load_checkpoint(args_opt.ckpt_file, net=gat_net)
+    load_checkpoint(args.ckpt_file, net=gat_net)
     gat_net.add_flags_recursive(fp16=True)
 
-    export(gat_net, Tensor(feature), Tensor(biases), file_name=args_opt.output_file, file_format="AIR")
+    export(gat_net, Tensor(feature), Tensor(biases), file_name=args.file_name, file_format=args.file_format)
diff --git a/model_zoo/official/gnn/gcn/export.py b/model_zoo/official/gnn/gcn/export.py
index e730f08c5b9..4d1a4296f46 100644
--- a/model_zoo/official/gnn/gcn/export.py
+++ b/model_zoo/official/gnn/gcn/export.py
@@ -16,24 +16,27 @@
 import argparse
 import numpy as np
 
-from mindspore import Tensor, context
-from mindspore.train.serialization import load_checkpoint, export
+from mindspore import Tensor, context, load_checkpoint, export
 
 from src.gcn import GCN
 from src.config import ConfigGCN
 
-context.set_context(mode=context.GRAPH_MODE, device_target="Ascend")
+parser = argparse.ArgumentParser(description="GCN export")
+parser.add_argument("--device_id", type=int, default=0, help="Device id")
+parser.add_argument("--ckpt_file", type=str, required=True, help="Checkpoint file path.")
+parser.add_argument("--dataset", type=str, default="cora", choices=["cora", "citeseer"], help="Dataset.")
+parser.add_argument("--file_name", type=str, default="gcn", help="output file name.")
+parser.add_argument("--file_format", type=str, choices=["AIR", "ONNX", "MINDIR"], default="AIR", help="file format")
+parser.add_argument("--device_target", type=str, default="Ascend",
+                    choices=["Ascend", "GPU", "CPU"], help="device target (default: Ascend)")
+args = parser.parse_args()
 
-if __name__ == '__main__':
-    parser = argparse.ArgumentParser(description='GCN_export')
-    parser.add_argument('--ckpt_file', type=str, default='', help='GCN ckpt file.')
-    parser.add_argument('--output_file', type=str, default='gcn.air', help='GCN output air name.')
-    parser.add_argument('--dataset', type=str, default='cora', help='GCN dataset name.')
-    args_opt = parser.parse_args()
+context.set_context(mode=context.GRAPH_MODE, device_target=args.device_target, device_id=args.device_id)
 
+if __name__ == "__main__":
     config = ConfigGCN()
 
-    if args_opt.dataset == "cora":
+    if args.dataset == "cora":
         input_dim = 1433
         class_num = 7
         adj = Tensor(np.zeros((2708, 2708), np.float64))
@@ -47,7 +50,7 @@ if __name__ == '__main__':
     gcn_net = GCN(config, input_dim, class_num)
 
     gcn_net.set_train(False)
-    load_checkpoint(args_opt.ckpt_file, net=gcn_net)
+    load_checkpoint(args.ckpt_file, net=gcn_net)
     gcn_net.add_flags_recursive(fp16=True)
 
-    export(gcn_net, adj, feature, file_name=args_opt.output_file, file_format="AIR")
+    export(gcn_net, adj, feature, file_name=args.file_name, file_format=args.file_format)
diff --git a/model_zoo/official/nlp/tinybert/export.py b/model_zoo/official/nlp/tinybert/export.py
index e02c82dca41..ba1d914c89c 100644
--- a/model_zoo/official/nlp/tinybert/export.py
+++ b/model_zoo/official/nlp/tinybert/export.py
@@ -25,11 +25,17 @@ from src.td_config import td_student_net_cfg
 from src.tinybert_model import BertModelCLS
 
 parser = argparse.ArgumentParser(description='tinybert task distill')
-parser.add_argument('--ckpt_file', type=str, required=True, help='tinybert ckpt file.')
-parser.add_argument('--output_file', type=str, default='tinybert', help='tinybert output air name.')
+parser.add_argument("--device_id", type=int, default=0, help="Device id")
+parser.add_argument("--ckpt_file", type=str, required=True, help="tinybert ckpt file.")
+parser.add_argument("--file_name", type=str, default="tinybert", help="output file name.")
+parser.add_argument("--file_format", type=str, choices=["AIR", "ONNX", "MINDIR"], default="AIR", help="file format")
+parser.add_argument("--device_target", type=str, default="Ascend",
+                    choices=["Ascend", "GPU", "CPU"], help="device target (default: Ascend)")
 parser.add_argument('--task_name', type=str, default='SST-2', choices=['SST-2', 'QNLI', 'MNLI'], help='task name')
 args = parser.parse_args()
 
+context.set_context(mode=context.GRAPH_MODE, device_target=args.device_target, device_id=args.device_id)
+
 DEFAULT_NUM_LABELS = 2
 DEFAULT_SEQ_LENGTH = 128
 DEFAULT_BS = 32
@@ -37,8 +43,6 @@ task_params = {"SST-2": {"num_labels": 2, "seq_length": 64},
                "QNLI": {"num_labels": 2, "seq_length": 128},
                "MNLI": {"num_labels": 3, "seq_length": 128}}
 
-context.set_context(mode=context.GRAPH_MODE, device_target="Ascend")
-
 class Task:
     """
     Encapsulation class of get the task parameter.
@@ -78,4 +82,5 @@ if __name__ == '__main__':
     token_type_id = Tensor(np.zeros((td_student_net_cfg.batch_size, task.seq_length), np.int32))
     input_mask = Tensor(np.zeros((td_student_net_cfg.batch_size, task.seq_length), np.int32))
 
-    export(eval_model, input_ids, token_type_id, input_mask, file_name=args.output_file, file_format="AIR")
+    input_data = [input_ids, token_type_id, input_mask]
+    export(eval_model, *input_data, file_name=args.file_name, file_format=args.file_format)
diff --git a/model_zoo/official/recommend/wide_and_deep/export.py b/model_zoo/official/recommend/wide_and_deep/export.py
index 0d322602638..58da4cb8899 100644
--- a/model_zoo/official/recommend/wide_and_deep/export.py
+++ b/model_zoo/official/recommend/wide_and_deep/export.py
@@ -13,48 +13,40 @@
 # limitations under the License.
 # ============================================================================
 """
-##############export checkpoint file into air and onnx models#################
+##############export checkpoint file into air, mindir and onnx models#################
 """
+import argparse
 import numpy as np
 
-from mindspore import Tensor, nn
-from mindspore.ops import operations as P
-from mindspore.train.serialization import load_checkpoint, load_param_into_net, export
+from mindspore import Tensor, context, load_checkpoint, export, load_param_into_net
 
-from src.wide_and_deep import WideDeepModel
+from eval import ModelBuilder
 from src.config import WideDeepConfig
 
-class PredictWithSigmoid(nn.Cell):
-    """
-    PredictWithSigmoid
-    """
-    def __init__(self, network):
-        super(PredictWithSigmoid, self).__init__()
-        self.network = network
-        self.sigmoid = P.Sigmoid()
+parser = argparse.ArgumentParser(description="wide_and_deep export")
+parser.add_argument("--device_id", type=int, default=0, help="Device id")
+parser.add_argument("--ckpt_file", type=str, required=True, help="Checkpoint file path.")
+parser.add_argument("--file_name", type=str, default="wide_and_deep", help="output file name.")
+parser.add_argument("--file_format", type=str, choices=["AIR", "ONNX", "MINDIR"], default="AIR", help="file format")
+parser.add_argument("--device_target", type=str, default="Ascend",
+                    choices=["Ascend", "GPU", "CPU"], help="device target (default: Ascend)")
+args = parser.parse_args()
 
-    def construct(self, batch_ids, batch_wts):
-        logits, _, = self.network(batch_ids, batch_wts)
-        pred_probs = self.sigmoid(logits)
-        return pred_probs
-
-def get_WideDeep_net(config):
-    """
-    Get network of wide&deep predict model.
-    """
-    WideDeep_net = WideDeepModel(config)
-    eval_net = PredictWithSigmoid(WideDeep_net)
-    return eval_net
+context.set_context(mode=context.GRAPH_MODE, device_target=args.device_target, device_id=args.device_id)
 
 if __name__ == '__main__':
     widedeep_config = WideDeepConfig()
     widedeep_config.argparse_init()
-    ckpt_path = widedeep_config.ckpt_path
-    net = get_WideDeep_net(widedeep_config)
-    param_dict = load_checkpoint(ckpt_path)
-    load_param_into_net(net, param_dict)
+
+    net_builder = ModelBuilder()
+    _, eval_net = net_builder.get_net(widedeep_config)
+
+    param_dict = load_checkpoint(args.ckpt_file)
+    load_param_into_net(eval_net, param_dict)
+    eval_net.set_train(False)
+
     ids = Tensor(np.ones([widedeep_config.eval_batch_size, widedeep_config.field_size]).astype(np.int32))
     wts = Tensor(np.ones([widedeep_config.eval_batch_size, widedeep_config.field_size]).astype(np.float32))
-    input_tensor_list = [ids, wts]
-    export(net, *input_tensor_list, file_name='wide_and_deep', file_format="ONNX")
-    export(net, *input_tensor_list, file_name='wide_and_deep', file_format="AIR")
+    label = Tensor(np.ones([widedeep_config.eval_batch_size, 1]).astype(np.float32))
+    input_tensor_list = [ids, wts, label]
+    export(eval_net, *input_tensor_list, file_name=args.file_name, file_format=args.file_format)