update Readme.md

Author: Messi-Q

Readme.md

@@ -5,7 +5,7 @@ We also present a automation tool of generate graph.
 
 ### Running project
 * To run program, use this command: python SmConVulDetector.py.
-* In addition, you can use specific hyperparameters to train the model. All the hyperparameters can be found in `parser.py`.
+* In addition, you can use specific hyper-parameters to train the model. All the hyper-parameters can be found in `parser.py`.
 
 Examples:
 ```shell
@@ -27,10 +27,15 @@ Then, you can find the training results in the `logs/`.
 ### Dataset
 Original smart contract source code:
 
-Ethereum smart contracts:  [Etherscan_contract](https://drive.google.com/open?id=1h9aFFSsL7mK4NmVJd4So7IJlFj9u0HRv)
+Ethereum smart contracts: [Etherscan_contract](https://drive.google.com/open?id=1h9aFFSsL7mK4NmVJd4So7IJlFj9u0HRv)
 
 Vntchain smart contacts: [Vntchain_contract](https://drive.google.com/open?id=1FTb__ERCOGNGM9dTeHLwAxBLw7X5Td4v)
 
 The train data after normalization:
 
 `data/LOOP_CORENODES_1317`, `LOOP_FULLNODES_1317`, `REENTRANCY_CORENODES_1671`, `REENTRANCY_FULLNODES_1671`
+
+
+### Reference
+1. The code borrows from [Graph_unet](https://github.com/bknyaz/graph_nn)
+2. Thomas N. Kipf, Max Welling, Semi-Supervised Classification with Graph Convolutional Networks, ICLR 2017

SmConVulDetector.py

@@ -15,7 +15,6 @@
 from models.gcn_modify import GCN_MODIFY
 from models.gcn_origin import GCN_ORIGIN
 from models.gat import GAT
-from models.mgcn import MGCN
 from sklearn import metrics
 
 print('using torch', torch.__version__)
@@ -211,15 +210,6 @@ def read_node_features(self, fpath, nodes, graphs, fn):
                            n_hidden=64,
                            n_class=loaders[0].dataset.num_classes,
                            dropout=args.dropout).to(args.device)
-    elif args.model == 'mgcn':
-        model = MGCN(in_features=loaders[0].dataset.num_features,
-                     out_features=loaders[0].dataset.num_classes,
-                     n_relations=2,
-                     n_hidden=args.n_hidden,
-                     filters=args.filters,
-                     dropout=args.dropout,
-                     adj_sq=args.adj_sq,
-                     scale_identity=args.scale_identity).to(args.device)
     elif args.model == 'gat':
         model = GAT(nfeat=loaders[0].dataset.num_features,
                     nhid=64,

models/mgcn.py

@@ -8,7 +8,7 @@ def parameter_parser():
                         choices=['REENTRANCY_CORENODES_1671', 'REENTRANCY_FULLNODES_1671',
                                  'LOOP_CORENODES_1317', 'LOOP_FULLNODES_1317'])
     parser.add_argument('-M', '--model', type=str, default='gcn_modify',
-                        choices=['gcn_modify', 'mgcn', 'gat', 'gcn_origin'])
+                        choices=['gcn_modify', 'gat', 'gcn_origin'])
     parser.add_argument('--lr', type=float, default=0.002, help='learning rate')
     parser.add_argument('--lr_decay_steps', type=str, default='1,3', help='learning rate')
     parser.add_argument('--wd', type=float, default=1e-4, help='weight decay')