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
|
import copy
import os
import torch
import torch.distributed as dist
import torch.nn.functional as F
from torch import Tensor
from torch.nn.parallel import DistributedDataParallel
from tqdm import tqdm
from torch_geometric.datasets import Reddit
from torch_geometric.loader import NeighborLoader
from torch_geometric.nn import SAGEConv
class SAGE(torch.nn.Module):
def __init__(
self,
in_channels: int,
hidden_channels: int,
out_channels: int,
num_layers: int = 2,
):
super().__init__()
self.convs = torch.nn.ModuleList()
self.convs.append(SAGEConv(in_channels, hidden_channels))
for _ in range(num_layers - 2):
self.convs.append(SAGEConv(hidden_channels, hidden_channels))
self.convs.append(SAGEConv(hidden_channels, out_channels))
def forward(self, x: Tensor, edge_index: Tensor) -> Tensor:
for i, conv in enumerate(self.convs):
x = conv(x, edge_index)
if i < len(self.convs) - 1:
x = x.relu_()
x = F.dropout(x, p=0.5, training=self.training)
return x
@torch.no_grad()
def inference(
self,
x_all: Tensor,
device: torch.device,
subgraph_loader: NeighborLoader,
) -> Tensor:
pbar = tqdm(total=len(subgraph_loader) * len(self.convs))
pbar.set_description('Evaluating')
# Compute representations of nodes layer by layer, using *all*
# available edges. This leads to faster computation in contrast to
# immediately computing the final representations of each batch:
for i, conv in enumerate(self.convs):
xs = []
for batch in subgraph_loader:
x = x_all[batch.node_id.to(x_all.device)].to(device)
x = conv(x, batch.edge_index.to(device))
x = x[:batch.batch_size]
if i < len(self.convs) - 1:
x = x.relu_()
xs.append(x.cpu())
pbar.update(1)
x_all = torch.cat(xs, dim=0)
pbar.close()
return x_all
def run(world_size: int, rank: int, local_rank: int):
# Will query the runtime environment for `MASTER_ADDR` and `MASTER_PORT`.
# Make sure, those are set!
dist.init_process_group('nccl', world_size=world_size, rank=rank)
# Download and unzip only with one process ...
if rank == 0:
dataset = Reddit('data/Reddit')
dist.barrier()
# ... and then read from all the other processes:
if rank != 0:
dataset = Reddit('data/Reddit')
dist.barrier()
data = dataset[0]
# Move to device for faster feature fetch.
data = data.to(local_rank, 'x', 'y')
# Split training indices into `world_size` many chunks:
train_idx = data.train_mask.nonzero(as_tuple=False).view(-1)
train_idx = train_idx.split(train_idx.size(0) // world_size)[rank]
kwargs = dict(batch_size=1024, num_workers=4, persistent_workers=True)
train_loader = NeighborLoader(
data,
input_nodes=train_idx,
num_neighbors=[25, 10],
shuffle=True,
drop_last=True,
**kwargs,
)
if rank == 0: # Create single-hop evaluation neighbor loader:
subgraph_loader = NeighborLoader(
copy.copy(data),
num_neighbors=[-1],
shuffle=False,
**kwargs,
)
# No need to maintain these features during evaluation:
del subgraph_loader.data.x, subgraph_loader.data.y
# Add global node index information:
subgraph_loader.data.node_id = torch.arange(data.num_nodes)
torch.manual_seed(12345)
model = SAGE(dataset.num_features, 256, dataset.num_classes).to(local_rank)
model = DistributedDataParallel(model, device_ids=[local_rank])
optimizer = torch.optim.Adam(model.parameters(), lr=0.001)
for epoch in range(1, 21):
model.train()
for batch in train_loader:
optimizer.zero_grad()
out = model(batch.x,
batch.edge_index.to(local_rank))[:batch.batch_size]
loss = F.cross_entropy(out, batch.y[:batch.batch_size])
loss.backward()
optimizer.step()
dist.barrier()
if rank == 0:
print(f'Epoch: {epoch:02d}, Loss: {loss:.4f}')
if rank == 0 and epoch % 5 == 0: # We evaluate on a single GPU for now
model.eval()
with torch.no_grad():
out = model.module.inference(
data.x,
local_rank,
subgraph_loader,
)
res = out.argmax(dim=-1) == data.y.to(out.device)
acc1 = int(res[data.train_mask].sum()) / int(data.train_mask.sum())
acc2 = int(res[data.val_mask].sum()) / int(data.val_mask.sum())
acc3 = int(res[data.test_mask].sum()) / int(data.test_mask.sum())
print(f'Train: {acc1:.4f}, Val: {acc2:.4f}, Test: {acc3:.4f}')
dist.barrier()
dist.destroy_process_group()
if __name__ == '__main__':
# Get the world size from the WORLD_SIZE variable or directly from SLURM:
world_size = int(
os.environ.get('WORLD_SIZE', os.environ.get('SLURM_NTASKS')))
# Likewise for RANK and LOCAL_RANK:
rank = int(os.environ.get('RANK', os.environ.get('SLURM_PROCID')))
local_rank = int(
os.environ.get('LOCAL_RANK', os.environ.get('SLURM_LOCALID')))
run(world_size, rank, local_rank)
|
