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
|
import os.path as osp
import time
import torch
import torch.nn.functional as F
import torch_geometric.transforms as T
from torch_geometric.datasets import Planetoid
from torch_geometric.nn import GCNConv
if torch.cuda.is_available():
device = torch.device('cuda')
elif hasattr(torch.backends, 'mps') and torch.backends.mps.is_available():
device = torch.device('mps')
else:
device = torch.device('cpu')
path = osp.dirname(osp.realpath(__file__))
path = osp.join(path, '..', '..', 'data', 'Planetoid')
dataset = Planetoid(
path, name='Cora', transform=T.Compose([
T.NormalizeFeatures(),
T.GCNNorm(),
]))
data = dataset[0].to(device)
class GCN(torch.nn.Module):
def __init__(self, in_channels, hidden_channels, out_channels):
super().__init__()
# Pre-process normalization to avoid CPU communication/graph breaks:
self.conv1 = GCNConv(in_channels, hidden_channels, normalize=False)
self.conv2 = GCNConv(hidden_channels, out_channels, normalize=False)
def forward(self, x, edge_index, edge_weight):
x = F.dropout(x, p=0.5, training=self.training)
x = self.conv1(x, edge_index, edge_weight).relu()
x = F.dropout(x, p=0.5, training=self.training)
x = self.conv2(x, edge_index)
return x
model = GCN(
in_channels=dataset.num_features,
hidden_channels=16,
out_channels=dataset.num_classes,
).to(device)
# Compile the model into an optimized version:
model = torch.compile(model, dynamic=False)
optimizer = torch.optim.Adam([
dict(params=model.conv1.parameters(), weight_decay=5e-4),
dict(params=model.conv2.parameters(), weight_decay=0)
], lr=0.01) # Only perform weight-decay on first convolution.
def train():
model.train()
optimizer.zero_grad()
out = model(data.x, data.edge_index, data.edge_weight)
loss = F.cross_entropy(out[data.train_mask], data.y[data.train_mask])
loss.backward()
optimizer.step()
return float(loss)
@torch.no_grad()
def test():
model.eval()
pred = model(data.x, data.edge_index, data.edge_weight).argmax(dim=-1)
accs = []
for mask in [data.train_mask, data.val_mask, data.test_mask]:
accs.append(int((pred[mask] == data.y[mask]).sum()) / int(mask.sum()))
return accs
times = []
for epoch in range(1, 201):
start = time.time()
loss = train()
train_acc, val_acc, test_acc = test()
times.append(time.time() - start)
print(f'Epoch: {epoch:03d}, Loss: {loss:.4f}, Train: {train_acc:.4f}, '
f'Val: {val_acc:.4f}, Test: {test_acc:.4f}')
print(f'Median time per epoch: {torch.tensor(times).median():.4f}s')
|
