diff --git a/pygod/detector/card.py b/pygod/detector/card.py
index d53fd37..6bfb5f7 100644
--- a/pygod/detector/card.py
+++ b/pygod/detector/card.py
@@ -65,6 +65,9 @@ class CARD(DeepDetector):
     gama: float, optional
         The proportion of the local reconstruction in contrastive learning module.
         Default: ``0.5``
+    alpha: float, optional
+        The proprotion of the community embedding in the conbine_encoder.
+        Default: ``0.1``
     verbose : int, optional
         Verbosity mode. Range in [0, 3]. Larger value for printing out
         more log information. Default: ``0``.
@@ -115,6 +118,7 @@ def __init__(self,
                  subgraph_num_neigh=4,
                  fp=0.6,
                  gama=0.5,
+                 alpha=0.1,
                  verbose=0,
                  save_emb=False,
                  compile_model=False,
@@ -138,6 +142,7 @@ def __init__(self,
         self.subgraph_num_neigh = subgraph_num_neigh
         self.fp = fp
         self.gama = gama
+        self.alpha = alpha
 
     def process_graph(self, data):
         community_adj, self.diff_data = CARDBase.process_graph(data)
@@ -151,14 +156,15 @@ def init_model(self, **kwargs):
                                    self.hid_dim)
 
         return CARDBase(in_dim=self.in_dim,
+                        subgraph_num_neigh=self.subgraph_num_neigh,
                         fp=self.fp,
                         gama=self.gama,
+                        alpha=self.alpha,
                         hid_dim=self.hid_dim,
                         num_layers=self.num_layers,
                         dropout=self.dropout,
                         act=self.act,
                         backbone=self.backbone,
-                        subgraph_num_neigh=self.subgraph_num_neigh,
                         **kwargs).to(self.device)
 
     def forward_model(self, data):
diff --git a/pygod/nn/card.py b/pygod/nn/card.py
index a1b9a37..ce6a8ef 100644
--- a/pygod/nn/card.py
+++ b/pygod/nn/card.py
@@ -36,6 +36,9 @@ class CARDBase(nn.Module):
     gama: float, optional
         The proportion of the local reconstruction in contrastive learning module.
         Default: ``0.5``
+    alpha: float, optional
+        The proprotion of the community embedding in the conbine_encoder.
+        Default: ``0.1``
     hid_dim :  int, optional
         Hidden dimension of model. Default: ``64``.
     num_layers : int, optional
@@ -54,9 +57,9 @@ class CARDBase(nn.Module):
 
     def __init__(self,
                  in_dim,
+                 subgraph_num_neigh=4,
                  fp=0.6,
                  gama=0.4,
-                 subgraph_num_neigh=4,
                  alpha=0.1,
                  hid_dim=64,
                  num_layers=4,
@@ -171,18 +174,18 @@ def loss_func(self, logits, diff_logits, x_, local_x_, x, con_label):
 
         Parameters
         ----------
-        logits : _type_
-            _description_
-        diff_logits : _type_
-            _description_
-        x_ : _type_
-            _description_
-        local_x_ : _type_
-            _description_
-        x : _type_
-            _description_
-        con_label : _type_
-            _description_
+        logits : torch.Tensor
+            Discriminator logits of positive subgraphs batch.
+        diff_logits : torch.Tensor
+            Discriminator logits of negative subgraphs batch.
+        x_ : torch.Tensor
+            Global reconstructed attribute embeddings.
+        local_x_ : torch.Tensor
+            Local reconstructed attribute embeddings.
+        x : torch.Tensor
+            Input attribute embeddings.
+        con_label : torch.Tensor
+            Contrastive learning pseudo label
 
         Returns
         -------
@@ -254,13 +257,6 @@ def _train_subgraph_network(self, data):
             x = subgraph.x
             edge_index = subgraph.edge_index
 
-            # diff_subgraphs = NeighborLoader(
-            #      self.diff, num_neighbors=[-1] * self.num_layers)
-            # diff_subgraph = diff_subgraphs([index])
-            # diff_subgraph.x[0, :] = 0
-            # diff_x = diff_subgraph.x.to(self.device)
-            # diff_edge_index = diff_subgraph.edge_index.to(self.device)
-
             ori_emb = self.encoder(x, edge_index)
             community_emb = self.community_encoder(community_adj)
             combine_emb = self.combine_encoder(