Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[BugFix] Fix a bug in using TUDataset #225

Merged
merged 2 commits into from
Apr 17, 2021
Merged

[BugFix] Fix a bug in using TUDataset #225

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
317fd52
Fix bugs in subgraph
THINK2TRY Apr 15, 2021
9f84044
Fix a bug in tu_dataset
THINK2TRY Apr 17, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions cogdl/data/data.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -648,12 +648,12 @@ def _sample_adj(self, batch_size, indices, indptr, size):

def csr_subgraph(self, node_idx):
indptr, indices, nodes, edges = subgraph_c(self._adj.row_ptr, self._adj.col, node_idx.cpu())
nodes = nodes.to(self._adj.device)
nodes_idx = node_idx.to(self._adj.device)
edge_weight = self.edge_weight[edges]

data = Graph(row_ptr=indptr, col=indices, weight=edge_weight)
for key in self.__keys__():
data[key] = self[key][nodes]
data[key] = self[key][nodes_idx]
return data

def subgraph(self, node_idx):
Expand Down
2 changes: 1 addition & 1 deletion cogdl/datasets/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -84,7 +84,7 @@ def build_dataset_from_path(data_path, task):
"ogbn-arxiv": "cogdl.datasets.ogb",
"ogbn-products": "cogdl.datasets.ogb",
"ogbn-proteins": "cogdl.datasets.ogb",
"ogbn-mag": "cogdl.datasets.pyg_ogb",
"ogbn-mag": "cogdl.datasets.ogb",
"ogbn-papers100M": "cogdl.datasets.ogb",
"ogbg-molbace": "cogdl.datasets.ogb",
"ogbg-molhiv": "cogdl.datasets.ogb",
Expand Down
2 changes: 1 addition & 1 deletion cogdl/datasets/tu_data.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -197,7 +197,7 @@ def read_tu_data(folder, prefix):
if x is not None:
x = x[:, num_node_attributes(x) :]
if edge_attr is not None:
edge_attr = edge_attr[:, num_edge_attributes(edge_attr)]
edge_attr = edge_attr[:, : num_edge_attributes(edge_attr)]

graphs = _split(edge_index, batch=batch, x=x, y=y, edge_attr=edge_attr)
return graphs, y
Expand Down
15 changes: 13 additions & 2 deletions cogdl/trainers/distributed_sampled_trainer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,6 +27,12 @@
from cogdl.trainers.base_trainer import BaseTrainer


import resource

rlimit = resource.getrlimit(resource.RLIMIT_NOFILE)
resource.setrlimit(resource.RLIMIT_NOFILE, (4096, rlimit[1]))


class DistributedSampledTrainer(BaseTrainer):
def __init__(self, args):
super(DistributedSampledTrainer, self).__init__(args)
Expand Down Expand Up @@ -274,6 +280,10 @@ def _test_step(self, split="val"):
return super(DistributedNeighborSamplerTrainer, self)._test_step()


def batcher(data):
return data[0]


@register_trainer("dist_saint")
class DistributedSAINTTrainer(DistributedSampledTrainer, SAINTTrainer):
@staticmethod
Expand Down Expand Up @@ -304,10 +314,11 @@ def build_dataloader(self, dataset, rank):
dataset=train_dataset,
batch_size=1,
shuffle=False,
num_workers=0,
num_workers=4,
persistent_workers=True,
pin_memory=True,
sampler=train_sampler,
collate_fn=lambda x: x[0],
collate_fn=batcher,
)

test_loader = NeighborSampler(
Expand Down