def loss(self, data: Graph, split="train"):
if split == "train":
mask = data.train_mask
elif split == "val":
mask = data.val_mask
else:
mask = data.test_mask
edge_index, edge_types = data.edge_index[:, mask], data.edge_attr[mask]
self.get_edge_set(edge_index, edge_types)
batch_edges, batch_attr, samples, rels, labels = sampling_edge_uniform(
edge_index,
edge_types,
self.edge_set,
self.sampling_rate,
self.num_rels,
label_smoothing=self.lbl_smooth,
num_entities=self.num_entities,
)
with data.local_graph():
data.edge_index = batch_edges
data.edge_attr = batch_attr
node_embed, rel_embed = self.forward(data)
sampled_nodes, reindexed_edges = torch.unique(samples, sorted=True, return_inverse=True)
assert (self.cache_index == sampled_nodes).any()
loss_n = self._loss(node_embed[reindexed_edges[0]], node_embed[reindexed_edges[1]], rel_embed[rels], labels)
loss_r = self.penalty * self._regularization([self.emb(sampled_nodes), rel_embed])
return loss_n + loss_r
def loss(self, data: Graph, scoring):
row, col = data.edge_index
edge_types = data.edge_attr
edge_index = torch.stack([row, col])
self.get_edge_set(edge_index, edge_types)
batch_edges, batch_attr, samples, rels, labels = sampling_edge_uniform(
(row, col),
edge_types,
self.edge_set,
self.sampling_rate,
self.num_rels,
label_smoothing=self.lbl_smooth,
num_entities=self.num_entities,
)
with data.local_graph():
data.edge_index = batch_edges
data.edge_attr = batch_attr
node_embed, rel_embed = self.forward(data)
sampled_nodes, reindexed_edges = torch.unique(samples,
sorted=True,
return_inverse=True)
assert (self.cache_index == sampled_nodes).any()
loss_n = self._loss(node_embed[reindexed_edges[0]],
node_embed[reindexed_edges[1]], rel_embed[rels],
labels, scoring)
loss_r = self.penalty * self._regularization(
[self.emb(sampled_nodes), rel_embed])
return loss_n + loss_r
def build_toy_data():
x = torch.randn(100, 10)
edge_index = torch.randint(0, 100, (2, 200))
g = Graph(x=x, edge_index=edge_index)
nedge = g.num_edges
edge_attr = torch.randn(nedge, 10)
g.edge_attr = edge_attr
return g
def read_triplet_data(folder):
filenames = ["train2id.txt", "valid2id.txt", "test2id.txt"]
count = 0
edge_index = []
edge_attr = []
count_list = []
triples = []
num_entities = 0
num_relations = 0
entity_dic = {}
relation_dic = {}
for filename in filenames:
with open(osp.join(folder, filename), "r") as f:
_ = int(f.readline().strip())
if "train" in filename:
train_start_idx = len(triples)
elif "valid" in filename:
valid_start_idx = len(triples)
elif "test" in filename:
test_start_idx = len(triples)
for line in f:
items = line.strip().split()
edge_index.append([int(items[0]), int(items[1])])
edge_attr.append(int(items[2]))
triples.append((int(items[0]), int(items[2]), int(items[1])))
if items[0] not in entity_dic:
entity_dic[items[0]] = num_entities
num_entities += 1
if items[1] not in entity_dic:
entity_dic[items[1]] = num_entities
num_entities += 1
if items[2] not in relation_dic:
relation_dic[items[2]] = num_relations
num_relations += 1
count += 1
count_list.append(count)
edge_index = torch.LongTensor(edge_index).t()
edge_attr = torch.LongTensor(edge_attr)
data = Graph()
data.edge_index = edge_index
data.edge_attr = edge_attr
def generate_mask(start, end):
mask = torch.BoolTensor(count)
mask[:] = False
mask[start:end] = True
return mask
data.train_mask = generate_mask(0, count_list[0])
data.val_mask = generate_mask(count_list[0], count_list[1])
data.test_mask = generate_mask(count_list[1], count_list[2])
return data, triples, train_start_idx, valid_start_idx, test_start_idx, num_entities, num_relations
