forked from BUPT-GAMMA/GammaGL
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathiehgcn_trainer.py
214 lines (178 loc) · 8.72 KB
/
iehgcn_trainer.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
202
203
204
205
206
207
208
209
210
import os
import sys
sys.path.insert(0, os.path.abspath('../../')) # adds path2gammagl to execute in command line.
# os.environ['TL_BACKEND'] = 'torch'
# os.environ['TL_BACKEND'] = 'paddle'
# os.environ['TL_BACKEND'] = 'mindspore' # unsupported
import argparse
import tensorlayerx as tlx
from gammagl.utils import add_self_loops, mask_to_index, degree, set_device
from tensorlayerx.model import TrainOneStep, WithLoss
import gammagl.transforms as T
from gammagl.datasets import HGBDataset, IMDB
from gammagl.models import ieHGCNModel
# This model only support dataset DBLP and IMDB.
targetType = {
'imdb': 'movie',
'dblp': 'author'
}
class SemiSpvzLoss(WithLoss):
def __init__(self, net, loss_fn):
super(SemiSpvzLoss, self).__init__(backbone=net, loss_fn=loss_fn)
def forward(self, data, y):
logits = self.backbone_network(data['x_dict'], data['edge_index_dict'], data['num_nodes_dict'])
train_logits = tlx.gather(logits[targetType[str.lower(args.dataset)]], data['train_idx'])
train_y = tlx.gather(data['y'], data['train_idx'])
loss = self._loss_fn(train_logits, train_y)
return loss
def calculate_acc(logits, y, metrics):
"""
Args:
logits: node logits
y: node labels
metrics: tensorlayerx.metrics
Returns:
rst
"""
metrics.update(logits, y)
rst = metrics.result()
metrics.reset()
return rst
def main(args):
if (str.lower(args.dataset) not in ['imdb', 'dblp']):
raise ValueError('Unknown dataset: {}'.format(args.dataset))
# load dataset
if str.lower(args.dataset) == 'imdb':
path = os.path.join(os.path.dirname(os.path.realpath(__file__)), '../IMDB')
metapaths = [[('movie', 'actor'), ('actor', 'movie')],
[('movie', 'director'), ('director', 'movie')]]
transform = T.AddMetaPaths(metapaths=metapaths, drop_orig_edges=True,
drop_unconnected_nodes=True)
dataset = IMDB(path, transform=transform)
graph = dataset[0]
y = graph[targetType[str.lower(args.dataset)]].y
num_classes = max(y) + 1
# for mindspore, it should be passed into node indices
train_idx = mask_to_index(graph[targetType[str.lower(args.dataset)]].train_mask)
test_idx = mask_to_index(graph[targetType[str.lower(args.dataset)]].test_mask)
val_idx = mask_to_index(graph[targetType[str.lower(args.dataset)]].val_mask)
num_nodes_dict = {targetType[str.lower(args.dataset)]: graph[targetType[str.lower(args.dataset)]].num_nodes}
else:
if tlx.BACKEND == 'tensorflow':
path = os.path.join(os.path.dirname(os.path.realpath(__file__)), '../DBLP')
metapaths = [[('author', 'paper'), ('paper', 'author')],
[('author', 'paper'), ('paper', 'term'), ('term', 'paper'), ('paper', 'author')],
[('author', 'paper'), ('paper', 'venue'), ('venue', 'paper'), ('paper', 'author')]]
transform = T.AddMetaPaths(metapaths=metapaths, drop_orig_edges=True,
drop_unconnected_nodes=True)
dataset = HGBDataset(path, args.dataset, transform=transform)
graph = dataset[0]
y = graph[targetType[str.lower(args.dataset)]].y
num_classes = (max(y) - min(y)) + 1
val_ratio = 0.2
train = mask_to_index(graph[targetType[str.lower(args.dataset)]].train_mask)
split = int(train.shape[0] * val_ratio)
train_idx = train[split:]
val_idx = train[:split]
test_idx = mask_to_index(graph[targetType[str.lower(args.dataset)]].test_mask)
num_nodes_dict = {targetType[str.lower(args.dataset)]: graph[targetType[str.lower(args.dataset)]].num_nodes}
else:
dataset = HGBDataset(args.dataset_path, args.dataset)
graph = dataset[0]
y = graph[targetType[str.lower(args.dataset)]].y
num_classes = (max(y) - min(y)) + 1
val_ratio = 0.2
train = mask_to_index(graph[targetType[str.lower(args.dataset)]].train_mask)
split = int(train.shape[0] * val_ratio)
train_idx = train[split:]
val_idx = train[:split]
test_idx = mask_to_index(graph[targetType[str.lower(args.dataset)]].test_mask)
num_nodes_dict = {'author': graph['author'].num_nodes, 'paper': graph['paper'].num_nodes,
'term': graph['term'].num_nodes, 'venue': graph['venue'].num_nodes}
if tlx.BACKEND == 'tensorflow':
edge_index_dict = graph.edge_index_dict
else:
edge_index_dict = {graph.edge_types[i]: graph.edge_stores[i]['edge_index'] for i in
range(len(graph.edge_stores))}
# for IMDB: train test val = 400, 3478, 400
# for DBLP: train test val = 974, 1420, 243
data = {
"x_dict": graph.x_dict,
"y": y,
"edge_index_dict": edge_index_dict,
"train_idx": train_idx,
"test_idx": test_idx,
"val_idx": val_idx,
"num_nodes_dict": num_nodes_dict,
}
net = ieHGCNModel(
num_layers=args.num_layers,
in_channels={node_type: node_shape.shape[1] for node_type, node_shape in graph.x_dict.items()},
hidden_channels=args.hidden_channels,
out_channels=num_classes,
attn_channels=args.attn_channels,
metadata=graph.metadata(),
batchnorm=False,
add_bias=True,
dropout_rate=args.dropout_rate,
name='iehgcn',
)
optimizer = tlx.optimizers.Adam(lr=args.lr, weight_decay=args.l2_coef)
metrics = tlx.metrics.Accuracy()
train_weights = net.trainable_weights
loss_func = tlx.losses.softmax_cross_entropy_with_logits
semi_spvz_loss = SemiSpvzLoss(net, loss_func)
train_one_step = TrainOneStep(semi_spvz_loss, optimizer, train_weights)
best_val_acc = 0
for epoch in range(args.n_epoch):
net.set_train()
train_loss = train_one_step(data, y)
net.set_eval()
logits = net(data['x_dict'], data['edge_index_dict'], data['num_nodes_dict'])
val_logits = tlx.gather(logits[targetType[str.lower(args.dataset)]], data['val_idx'])
val_y = tlx.gather(data['y'], data['val_idx'])
val_acc = calculate_acc(val_logits, val_y, metrics)
print("Epoch [{:0>3d}] ".format(epoch + 1)
+ " train_loss: {:.4f}".format(train_loss.item())
+ " val_acc: {:.4f}".format(val_acc))
# save best model on evaluation set
if val_acc > best_val_acc:
best_val_acc = val_acc
net.save_weights(args.best_model_path + net.name + ".npz", format='npz_dict')
net.load_weights(args.best_model_path + net.name + ".npz", format='npz_dict')
net.set_eval()
logits = net(data['x_dict'], data['edge_index_dict'], data['num_nodes_dict'])
test_logits = tlx.gather(logits[targetType[str.lower(args.dataset)]], data['test_idx'])
test_y = tlx.gather(data['y'], data['test_idx'])
if tlx.BACKEND == 'torch':
test_logits = test_logits.detach().numpy() # torch
else:
test_logits = test_logits.numpy()
import numpy as np
test_logits = np.argmax(test_logits, axis=1)
if tlx.BACKEND == 'torch':
test_y = test_y.detach().numpy() # torch
else:
test_y = test_y.numpy()
from sklearn.metrics import f1_score
macro_f1 = f1_score(y_true=test_y, y_pred=test_logits, average='macro')
micro_f1 = f1_score(y_true=test_y, y_pred=test_logits, average='micro')
print("Macro-F1: {:.4f}".format(macro_f1))
print("Micro-F1: {:.4f}".format(micro_f1))
return macro_f1
if __name__ == '__main__':
# parameters setting
parser = argparse.ArgumentParser()
parser.add_argument("--dataset_path", type=str, default=r'../', help="path to save dataset, not work")
parser.add_argument('--dataset', type=str, default='DBLP', help='dataset, IMDB or DBLP')
parser.add_argument("--lr", type=float, default=0.01, help="learning rate")
parser.add_argument("--n_epoch", type=int, default=30, help="number of epoch")
parser.add_argument("--num_layers", type=int, default=4, help="number of layers")
parser.add_argument("--hidden_channels", type=int, default=[64, 32, 16], help="dimention of hidden layers")
parser.add_argument("--attn_channels", type=int, default=32, help="dimention of attention layers")
parser.add_argument("--l2_coef", type=float, default=0.0005, help="l2 loss coeficient")
parser.add_argument("--dropout_rate", type=float, default=0.1, help="dropout_rate")
parser.add_argument("--gpu", type=int, default=0, help="gpu id")
parser.add_argument("--best_model_path", type=str, default=r'./', help="path to save best model")
args = parser.parse_args()
main(args)