diff --git a/autogl/module/nas/algorithm/__init__.py b/autogl/module/nas/algorithm/__init__.py
index 612102a..537f223 100644
--- a/autogl/module/nas/algorithm/__init__.py
+++ b/autogl/module/nas/algorithm/__init__.py
@@ -4,5 +4,6 @@ NAS algorithms
 
 from .base import BaseNAS
 from .darts import Darts
+from .enas import Enas
 
-__all__ = ["BaseNAS", "Darts"]
+__all__ = ["BaseNAS", "Darts", "Enas"]
diff --git a/autogl/module/nas/enas.py b/autogl/module/nas/algorithm/enas.py
similarity index 97%
rename from autogl/module/nas/enas.py
rename to autogl/module/nas/algorithm/enas.py
index 2cad798..3d175aa 100644
--- a/autogl/module/nas/enas.py
+++ b/autogl/module/nas/algorithm/enas.py
@@ -6,9 +6,9 @@ import torch
 import torch.nn as nn
 import torch.nn.functional as F
 
-from .nas import BaseNAS
-from .space import SpaceModel
-from .utils import AverageMeterGroup, replace_layer_choice, replace_input_choice
+from .base import BaseNAS
+from ..space import BaseSpace
+from ..utils import AverageMeterGroup, replace_layer_choice, replace_input_choice
 from nni.nas.pytorch.fixed import apply_fixed_architecture
 _logger = logging.getLogger(__name__)
 def _get_mask(sampled, total):
@@ -288,10 +288,10 @@ class Enas(BaseNAS):
         self.ctrl_kwargs=ctrl_kwargs
         self.ctrl_lr=ctrl_lr
 
-    def search(self, space, dset, trainer):
+    def search(self, space: BaseSpace, dset, estimator):
         self.model = space
         self.dataset = dset#.to(self.device)
-        self.trainer = trainer
+        self.estimator = estimator
         self.model_optim = torch.optim.SGD(
             self.model.parameters(), lr=0.01, weight_decay=3e-4
         )
@@ -313,7 +313,8 @@ class Enas(BaseNAS):
             self._train_controller(i)
         
         selection=self.export()
-        return SpaceModel(space,selection,self.device)
+        return space.export(selection,self.device)
+
     def _train_model(self, epoch): 
         self.model.train()
         self.controller.eval()
@@ -364,5 +365,5 @@ class Enas(BaseNAS):
             return self.controller.resample()
 
     def _infer(self):
-        metric, loss = self.trainer.infer(self.model, self.dataset)
+        metric, loss = self.estimator.infer(self.model, self.dataset)
         return metric, loss
diff --git a/examples/test_nas.py b/examples/test_nas.py
index 22aa03c..693cd10 100644
--- a/examples/test_nas.py
+++ b/examples/test_nas.py
@@ -8,6 +8,7 @@ from autogl.module.model import BaseModel
 from autogl.module.train import NodeClassificationFullTrainer
 from autogl.module.nas import Darts, OneShotEstimator, SinglePathNodeClassificationSpace
 from autogl.module.train import Acc
+from autogl.module.nas.algorithm.enas import Enas
 
 if __name__ == '__main__':
     dataset = build_dataset_from_name('cora')
@@ -30,6 +31,7 @@ if __name__ == '__main__':
             feval=['acc'],
             loss="nll_loss",
             lr_scheduler_type=None,),
+        #nas_algorithms=[Enas()],
         nas_algorithms=[Darts(num_epochs=1)],
         nas_spaces=[SinglePathNodeClassificationSpace(hidden_dim=16, ops=[GCNConv, GCNConv])],
         nas_estimators=[OneShotEstimator()]