-
Notifications
You must be signed in to change notification settings - Fork 3
/
model_config.py
201 lines (191 loc) · 6.3 KB
/
model_config.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
def set_model_config(args):
if args['model'] == 'gin_supervised_contextpred':
config = {
"batch_size": 128,
"jk": "concat",
"lr": 0.001,
"patience": 30,
"readout": "sum",
"weight_decay": 0.001
}
elif args['model'] == 'GCN':
config = {
"batch_size": 128,
"batchnorm": False,
"dropout": 0.1,
"gnn_hidden_feats": 128,
"lr": 0.002,
"num_gnn_layers": 2,
"patience": 30,
"predictor_hidden_feats": 64,
"residual": True,
"weight_decay": 0.001
}
elif args['model'] == 'GAT':
config = {
"alpha": 0.5,
"batch_size": 128,
"dropout": 0.05,
"gnn_hidden_feats": 128,
"lr": 0.01,
"num_gnn_layers": 2,
"num_heads": 6,
"patience": 30,
"predictor_hidden_feats": 128,
"residual": False,
"weight_decay": 0.0005
}
elif args['model'] == 'MPNN':
config = {
"batch_size": 128,
"edge_hidden_feats": 64,
"lr": 0.001,
"node_out_feats": 48,
"num_layer_set2set": 2,
"num_step_message_passing": 2,
"num_step_set2set": 2,
"patience": 30,
"weight_decay": 0.0005
}
elif args['model'] == 'AttentiveFP':
config = {
"batch_size": 128,
"dropout": 0.2,
"graph_feat_size": 32,
"lr": 0.01,
"num_layers": 2,
"num_timesteps": 3,
"patience": 30,
"weight_decay": 0.001
}
return config
def GCN_config(args, config):
if args['dataset'] == 'FreeSolv':
config['lr'] = 0.001
config['batch_size'] = 64
config['weight_decay'] = 0.0
elif args['dataset'] == 'ESOL':
config['lr'] = 0.001
config['batch_size'] = 64
config['weight_decay'] = 0.0
elif args['dataset'] == 'Lipophilicity':
config['lr'] = 0.001
config['batch_size'] = 64
config['weight_decay'] = 0.0
elif args['dataset'] == 'BBBP':
config['lr'] = 0.001
config['batch_size'] = 256
config['weight_decay'] = 0.0005
elif args['dataset'] == 'BACE':
config['lr'] = 0.001
config['batch_size'] = 256
config['weight_decay'] = 0.0005
elif args['dataset'] == 'HIV':
config['lr'] = 0.004
config['batch_size'] = config['batch_size']
config['weight_decay'] = config['weight_decay']
return config
def GAT_config(args, config):
if args['dataset'] == 'FreeSolv':
config['lr'] = 0.001
config['batch_size'] = 256
config['weight_decay'] = 0.0005
elif args['dataset'] == 'ESOL':
config['lr'] = 0.001
config['batch_size'] = 256
config['weight_decay'] = 0.0005
elif args['dataset'] == 'Lipophilicity':
config['lr'] = 0.0005
config['batch_size'] = 64
config['weight_decay'] = 0.0
elif args['dataset'] == 'BBBP':
config['lr'] = 0.0005
config['batch_size'] = 64
config['weight_decay'] = 0.0
elif args['dataset'] == 'BACE':
config['lr'] = 0.001
config['batch_size'] = 64
config['weight_decay'] = 0.0
elif args['dataset'] == 'HIV':
config['lr'] = 0.004
config['batch_size'] = config['batch_size']
config['weight_decay'] = config['weight_decay']
return config
def MPNN_config(args, config):
if args['dataset'] == 'FreeSolv':
config['lr'] = 0.005
config['batch_size'] = 128
config['weight_decay'] = 0.0005
elif args['dataset'] == 'ESOL':
config['lr'] = 0.001
config['batch_size'] = 64
config['weight_decay'] = 0.0
elif args['dataset'] == 'Lipophilicity':
config['lr'] = 0.001
config['batch_size'] = 64
config['weight_decay'] = 0.0005
elif args['dataset'] == 'BBBP':
config['lr'] = 0.005
config['batch_size'] = 64
config['weight_decay'] = 0.0005
elif args['dataset'] == 'BACE':
config['lr'] = 0.0005
config['batch_size'] = 64
config['weight_decay'] = 0.0005
elif args['dataset'] == 'HIV':
config['lr'] = config['batch_size']
config['batch_size'] = config['batch_size']
config['weight_decay'] = config['weight_decay']
return config
def AttentiveFP_config(args, config):
if args['dataset'] == 'FreeSolv':
config['lr'] = 0.01
config['batch_size'] = 64
config['weight_decay'] = 0.0005
elif args['dataset'] == 'ESOL':
config['lr'] = 0.005
config['batch_size'] = 64
config['weight_decay'] = 0.0
elif args['dataset'] == 'Lipophilicity':
config['lr'] = 0.005
config['batch_size'] = 64
config['weight_decay'] = 0.0
elif args['dataset'] == 'BBBP':
config['lr'] = 0.005
config['batch_size'] = 64
config['weight_decay'] = 0.0
elif args['dataset'] == 'BACE':
config['lr'] = 0.005
config['batch_size'] = 64
config['weight_decay'] = 0.0005
elif args['dataset'] == 'HIV':
config['lr'] = config['batch_size']
config['batch_size'] = config['batch_size']
config['weight_decay'] = config['weight_decay']
return config
def Pretrained_GIN_config(args, config):
if args['dataset'] == 'FreeSolv':
config['lr'] = 0.001
config['batch_size'] = 256
config['weight_decay'] = 0.001
elif args['dataset'] == 'ESOL':
config['lr'] = 0.0005
config['batch_size'] = 256
config['weight_decay'] = 0.001
elif args['dataset'] == 'Lipophilicity':
config['lr'] = 0.0005
config['batch_size'] = 128
config['weight_decay'] = 0.0
elif args['dataset'] == 'BBBP':
config['lr'] = 0.0005
config['batch_size'] = 64
config['weight_decay'] = 0.0005
elif args['dataset'] == 'BACE':
config['lr'] = 0.001
config['batch_size'] = 128
config['weight_decay'] = 0.0
elif args['dataset'] == 'HIV':
config['lr'] = config['batch_size']
config['batch_size'] = config['batch_size']
config['weight_decay'] = config['weight_decay']
return config