-
Notifications
You must be signed in to change notification settings - Fork 456
/
Copy pathtrain_COLLAB_edge_classification.py
executable file
·155 lines (123 loc) · 5.21 KB
/
train_COLLAB_edge_classification.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
"""
Utility functions for training one epoch
and evaluating one epoch
"""
import torch
import torch.nn as nn
from torch.utils.data import DataLoader
import dgl
import numpy as np
from tqdm import tqdm
"""
For GCNs
"""
def train_epoch_sparse(model, optimizer, device, graph, train_edges, batch_size, epoch, monet_pseudo=None):
model.train()
train_edges = train_edges.to(device)
total_loss = total_examples = 0
for perm in tqdm(DataLoader(range(train_edges.size(0)), batch_size, shuffle=True)):
optimizer.zero_grad()
graph = graph.to(device)
x = graph.ndata['feat'].to(device)
e = graph.edata['feat'].to(device).float()
if monet_pseudo is not None:
# Assign e as pre-computed pesudo edges for MoNet
e = monet_pseudo.to(device)
# Compute node embeddings
try:
x_pos_enc = graph.ndata['pos_enc'].to(device)
sign_flip = torch.rand(x_pos_enc.size(1)).to(device)
sign_flip[sign_flip>=0.5] = 1.0; sign_flip[sign_flip<0.5] = -1.0
x_pos_enc = x_pos_enc * sign_flip.unsqueeze(0)
h = model(graph, x, e, x_pos_enc)
except:
h = model(graph, x, e)
# Positive samples
edge = train_edges[perm].t()
pos_out = model.edge_predictor( h[edge[0]], h[edge[1]] )
# Just do some trivial random sampling
edge = torch.randint(0, x.size(0), edge.size(), dtype=torch.long, device=x.device)
neg_out = model.edge_predictor( h[edge[0]], h[edge[1]] )
loss = model.loss(pos_out, neg_out)
loss.backward()
optimizer.step()
num_examples = pos_out.size(0)
total_loss += loss.detach().item() * num_examples
total_examples += num_examples
return total_loss/total_examples, optimizer
def evaluate_network_sparse(model, device, graph, pos_train_edges,
pos_valid_edges, neg_valid_edges,
pos_test_edges, neg_test_edges,
evaluator, batch_size, epoch, monet_pseudo=None):
model.eval()
with torch.no_grad():
graph = graph.to(device)
x = graph.ndata['feat'].to(device)
e = graph.edata['feat'].to(device).float()
if monet_pseudo is not None:
# Assign e as pre-computed pesudo edges for MoNet
e = monet_pseudo.to(device)
# Compute node embeddings
try:
x_pos_enc = graph.ndata['pos_enc'].to(device)
h = model(graph, x, e, x_pos_enc)
except:
h = model(graph, x, e)
pos_train_edges = pos_train_edges.to(device)
pos_valid_edges = pos_valid_edges.to(device)
neg_valid_edges = neg_valid_edges.to(device)
pos_test_edges = pos_test_edges.to(device)
neg_test_edges = neg_test_edges.to(device)
pos_train_preds = []
for perm in DataLoader(range(pos_train_edges.size(0)), batch_size):
edge = pos_train_edges[perm].t()
pos_train_preds += [model.edge_predictor(h[edge[0]], h[edge[1]]).squeeze().cpu()]
pos_train_pred = torch.cat(pos_train_preds, dim=0)
pos_valid_preds = []
for perm in DataLoader(range(pos_valid_edges.size(0)), batch_size):
edge = pos_valid_edges[perm].t()
pos_valid_preds += [model.edge_predictor(h[edge[0]], h[edge[1]]).squeeze().cpu()]
pos_valid_pred = torch.cat(pos_valid_preds, dim=0)
neg_valid_preds = []
for perm in DataLoader(range(pos_valid_edges.size(0)), batch_size):
edge = neg_valid_edges[perm].t()
neg_valid_preds += [model.edge_predictor(h[edge[0]], h[edge[1]]).squeeze().cpu()]
neg_valid_pred = torch.cat(neg_valid_preds, dim=0)
pos_test_preds = []
for perm in DataLoader(range(pos_test_edges.size(0)), batch_size):
edge = pos_test_edges[perm].t()
pos_test_preds += [model.edge_predictor(h[edge[0]], h[edge[1]]).squeeze().cpu()]
pos_test_pred = torch.cat(pos_test_preds, dim=0)
neg_test_preds = []
for perm in DataLoader(range(pos_test_edges.size(0)), batch_size):
edge = neg_test_edges[perm].t()
neg_test_preds += [model.edge_predictor(h[edge[0]], h[edge[1]]).squeeze().cpu()]
neg_test_pred = torch.cat(neg_test_preds, dim=0)
train_hits = []
for K in [10, 50, 100]:
evaluator.K = K
train_hits.append(
evaluator.eval({
'y_pred_pos': pos_train_pred,
'y_pred_neg': neg_valid_pred, # negative samples for valid == training
})[f'hits@{K}']
)
valid_hits = []
for K in [10, 50, 100]:
evaluator.K = K
valid_hits.append(
evaluator.eval({
'y_pred_pos': pos_valid_pred,
'y_pred_neg': neg_valid_pred,
})[f'hits@{K}']
)
test_hits = []
for K in [10, 50, 100]:
evaluator.K = K
test_hits.append(
evaluator.eval({
'y_pred_pos': pos_test_pred,
'y_pred_neg': neg_test_pred,
})[f'hits@{K}']
)
return train_hits, valid_hits, test_hits