def track_time(graph_name, format, k):
device = utils.get_bench_device()
graph = utils.get_graph(graph_name, format)
graph = graph.to(device)
graph = graph.formats([format])
# dry run
dgl.khop_graph(graph, k)
# timing
with utils.Timer() as t:
for i in range(10):
gg = dgl.khop_graph(graph, k)
return t.elapsed_secs / 10
def track_time(graph_name, format, k):
device = utils.get_bench_device()
graph = utils.get_graph(graph_name, format)
graph = graph.to(device)
graph = graph.formats([format])
# dry run
dgl.khop_graph(graph, k)
# timing
t0 = time.time()
for i in range(10):
gg = dgl.khop_graph(graph, k)
t1 = time.time()
return (t1 - t0) / 10
def calculate_homophily(g,
labels,
K=1,
method="edge",
multilabels=False,
heterograph=False):
assert method in ["edge", "node"]
if multilabels:
assert len(labels.shape) == 2
else:
if (labels.max() == 1) and len(labels.shape) > 1:
labels = labels.argmax(dim=1)
if heterograph:
target_mask = g.ndata['target_mask']
target_ids = g.ndata[dgl.NID][target_mask]
num_target = target_mask.sum().item()
g = g.subgraph(np.arange(g.number_of_nodes())[target_mask])
g = dgl.khop_graph(g, K)
src, dst = g.edges()
# if multilabels:
# out = 0
# for c in labels.shape[1]:
# mask = (labels[src, c])
mask = (labels[src] == labels[dst]).float()
if method == "edge":
out = mask.mean(dim=0)
elif method == "node":
g.edata["mask"] = mask
g.update_all(fn.copy_e("mask", "m"), fn.mean("m", "out"))
out = g.ndata.pop("out").mean(dim=0)
# for multilabels, we average homophily across labels
return out.mean(0).item()
def _test(g):
for k in range(4):
g_k = dgl.khop_graph(g, k)
# use original graph to do message passing for k times.
g.ndata['h'] = feat
for _ in range(k):
g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
h_0 = g.ndata.pop('h')
# use k-hop graph to do message passing for one time.
g_k.ndata['h'] = feat
g_k.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
h_1 = g_k.ndata.pop('h')
assert F.allclose(h_0, h_1, rtol=1e-3, atol=1e-3)
def test_khop_graph():
N = 20
feat = F.randn((N, 5))
g = dgl.DGLGraph(nx.erdos_renyi_graph(N, 0.3))
for k in range(4):
g_k = dgl.khop_graph(g, k)
# use original graph to do message passing for k times.
g.ndata['h'] = feat
for _ in range(k):
g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
h_0 = g.ndata.pop('h')
# use k-hop graph to do message passing for one time.
g_k.ndata['h'] = feat
g_k.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
h_1 = g_k.ndata.pop('h')
assert F.allclose(h_0, h_1, rtol=1e-3, atol=1e-3)
logits = model(g, features)
logits = logits[mask]
labels = labels[mask]
_, indices = th.max(logits, dim=1)
correct = th.sum(indices == labels)
return correct.item() * 1.0 / len(labels)
# 训练网络
import time
import numpy as np
# g, features, labels, train_mask, test_mask = load_cora_data()
g = dgl.graph(([0, 1, 2, 3, 4, 0, 1, 2, 3, 4], [0, 1, 2, 3, 4, 1, 2, 3, 4, 0]))
g = dgl.khop_graph(g, 3)
input = th.ones(32, 5, 1)
net = Net(1, 5)
res = net(g, input)
pass
# # 定义优化器
# optimizer = th.optim.Adam(net.parameters(), lr=1e-3)
# dur = [] # 时间
# for epoch in range(100):
# print(epoch)
# if epoch >= 3:
# t0 = time.time()
# net.train()
# logits = net(g, features)
# logp = F.log_softmax(logits, 1)
