Skip to content

Category: B1 (Bonus); Team name: DLLB; Dataset: Reddit #231

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

Open
dleko11 wants to merge 12 commits into
base: main
Choose a base branch
from
Open

Category: B1 (Bonus); Team name: DLLB; Dataset: Reddit #231

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
12 commits
Select commit Hold shift + click to select a range
205cadc
Category: B1;  Team name: DLLB;  Dataset: FakeDataset
Nov 3, 2025
736948e
Removed some prints.
luka-benic Nov 3, 2025
c9cf909
Resolved conflicts in split_utils
Nov 4, 2025
b98ada9
Added unit tests; fixed docstrings.
Nov 19, 2025
43997c3
Fixed memory_plotting output path.
Nov 19, 2025
6d50b4d
Improved plotting utilities.
Nov 19, 2025
1a12c67
Improved plotting.
Nov 19, 2025
4ed754c
B1 Bonus.
Nov 19, 2025
24a3c8c
Merge branch 'main' into on_disk_transductive
dleko11 Nov 19, 2025
ce9d911
Added some test; fixed some issues.
Nov 20, 2025
1941aed
Added additional models to the tutorial.
Nov 21, 2025
b55b876
Fixed minor path issues.
Nov 21, 2025
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: 4 additions & 0 deletions .gitignore
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -186,6 +186,10 @@ notebooks/tmp
/tutorials/lightning_logs/
/tutorials/datasets/

# Sweep outputs
/tools/sweep_tools/outputs
/tools/memory_usage_tracking/outputs

# wandb
wandb/

Expand Down
55 changes: 55 additions & 0 deletions configs/dataset/graph/cocitation_citeseer_for_partitioning.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,55 @@
# Dataset loader config
loader:
_target_: topobench.data.loaders.PlanetoidDatasetLoader
parameters:
data_domain: graph
data_type: cocitation_on_disk
data_name: citeseer
data_dir: ${paths.data_dir}/${dataset.loader.parameters.data_domain}/${dataset.loader.parameters.data_type}

# Choose memory mode
memory_type: on_disk_cluster # "in_memory", "on_disk", or "on_disk_cluster"

# Global partition settings
cluster:
num_parts: 32 # number of clusters for single global partition
recursive: false # forwarded to PyG ClusterData
keep_inter_cluster_edges: false # standard Cluster-GCN style
sparse_format: csr # required by our block-stream loader

# Streaming / loader settings for block-wise training
stream:
q: 4 # clusters per batch (Cluster-GCN "bsize")
num_workers: 0 # increase if needed
pin_memory: false # true if you want faster H2D
with_edge_attr: false # WebKB has no edge_attr
precompute_split_parts: true # write parts_with_{train,val,test}.npy

# How to store features on disk
dtype_policy: float32 # current ClusterOnDisk uses float32

# Dataset parameters (unchanged)
parameters:
num_features: 3703
num_classes: 6
num_nodes: 3327
task: classification
loss_type: cross_entropy
monitor_metric: accuracy
task_level: node

# Splits (unchanged; used by pack_global_partition)
split_params:
learning_setting: transductive # transductive # inductive
data_split_dir: ${paths.data_dir}/data_splits/${dataset.loader.parameters.data_name}
data_seed: 0
split_type: random
k: 10
train_prop: 0.5

# Dataloader parameters for legacy paths (kept; ignored in on_disk_cluster mode)
dataloader_params:
batch_size: 1
drop_last: True
num_workers: 0
pin_memory: False
55 changes: 55 additions & 0 deletions configs/dataset/graph/cocitation_citeseer_on_disk.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,55 @@
# Dataset loader config
loader:
_target_: topobench.data.loaders.OnDiskDatasetLoader
parameters:
data_domain: graph
data_type: cocitation_on_disk
data_name: citeseer_on_disk
data_dir: ${paths.data_dir}/${dataset.loader.parameters.data_domain}/${dataset.loader.parameters.data_type}

# Choose memory mode
memory_type: on_disk # "in_memory", "on_disk", or "on_disk_cluster"

# Global partition setting
cluster:
num_parts: 32 # number of clusters for single global partition
recursive: false # forwarded to PyG ClusterData
keep_inter_cluster_edges: false # standard Cluster-GCN style
sparse_format: csr # required by our block-stream loader

# Streaming / loader settings for block-wise training
stream:
q: 4 # clusters per batch (Cluster-GCN "bsize")
num_workers: 0 # increase if needed
pin_memory: false # true if you want faster H2D
with_edge_attr: false # WebKB has no edge_attr
precompute_split_parts: true # write parts_with_{train,val,test}.npy

# How to store features on disk
dtype_policy: float32 # current ClusterOnDisk uses float32

# Dataset parameters (unchanged)
parameters:
num_features: 3703
num_classes: 6
num_nodes: 3327
task: classification
loss_type: cross_entropy
monitor_metric: accuracy
task_level: node

# Splits (unchanged; used by pack_global_partition)
split_params:
learning_setting: transductive # transductive # inductive
data_split_dir: ${paths.data_dir}/data_splits/${dataset.loader.parameters.data_name}
data_seed: 0
split_type: random
k: 10
train_prop: 0.5

# Dataloader parameters for legacy paths (kept; ignored in on_disk_cluster mode)
dataloader_params:
batch_size: 1
drop_last: True
num_workers: 0
pin_memory: False
55 changes: 55 additions & 0 deletions configs/dataset/graph/cocitation_cora_for_partitioning.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,55 @@
# Dataset loader config
loader:
_target_: topobench.data.loaders.PlanetoidDatasetLoader
parameters:
data_domain: graph
data_type: cocitation_on_disk
data_name: Cora
data_dir: ${paths.data_dir}/${dataset.loader.parameters.data_domain}/${dataset.loader.parameters.data_type}

# Choose memory mode
memory_type: on_disk_cluster # "in_memory", "on_disk", or "on_disk_cluster"

# Global partition settings
cluster:
num_parts: 32 # number of clusters for single global partition
recursive: false # forwarded to PyG ClusterData
keep_inter_cluster_edges: false # standard Cluster-GCN style
sparse_format: csr # required by our block-stream loader

# Streaming / loader settings for block-wise training
stream:
q: 4 # clusters per batch (Cluster-GCN "bsize")
num_workers: 0 # increase if needed
pin_memory: false # true if you want faster H2D
with_edge_attr: false # WebKB has no edge_attr
precompute_split_parts: true # write parts_with_{train,val,test}.npy

# How to store features on disk
dtype_policy: float32 # current ClusterOnDisk uses float32

# Dataset parameters (unchanged)
parameters:
num_features: 1433
num_classes: 7
num_nodes: 2708
task: classification
loss_type: cross_entropy
monitor_metric: accuracy
task_level: node

# Splits (unchanged; used by pack_global_partition)
split_params:
learning_setting: transductive # transductive # inductive
data_split_dir: ${paths.data_dir}/data_splits/${dataset.loader.parameters.data_name}
data_seed: 0
split_type: random
k: 10
train_prop: 0.5

# Dataloader parameters for legacy paths (kept; ignored in on_disk_cluster mode)
dataloader_params:
batch_size: 1
drop_last: True
num_workers: 0
pin_memory: False
55 changes: 55 additions & 0 deletions configs/dataset/graph/cocitation_cora_on_disk.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,55 @@
# Dataset loader config
loader:
_target_: topobench.data.loaders.OnDiskDatasetLoader
parameters:
data_domain: graph
data_type: cocitation_on_disk
data_name: Cora_on_disk
data_dir: ${paths.data_dir}/${dataset.loader.parameters.data_domain}/${dataset.loader.parameters.data_type}

# Choose memory mode
memory_type: on_disk # "in_memory", "on_disk", or "on_disk_cluster"

# Global partition setting
cluster:
num_parts: 32 # number of clusters for single global partition
recursive: false # forwarded to PyG ClusterData
keep_inter_cluster_edges: false # standard Cluster-GCN style
sparse_format: csr # required by our block-stream loader

# Streaming / loader settings for block-wise training
stream:
q: 4 # clusters per batch (Cluster-GCN "bsize")
num_workers: 0 # increase if needed
pin_memory: false # true if you want faster H2D
with_edge_attr: false # WebKB has no edge_attr
precompute_split_parts: true # write parts_with_{train,val,test}.npy

# How to store features on disk
dtype_policy: float32 # current ClusterOnDisk uses float32

# Dataset parameters (unchanged)
parameters:
num_features: 1433
num_classes: 7
num_nodes: 2708
task: classification
loss_type: cross_entropy
monitor_metric: accuracy
task_level: node

# Splits (unchanged; used by pack_global_partition)
split_params:
learning_setting: transductive # transductive # inductive
data_split_dir: ${paths.data_dir}/data_splits/${dataset.loader.parameters.data_name}
data_seed: 0
split_type: random
k: 10
train_prop: 0.5

# Dataloader parameters for legacy paths (kept; ignored in on_disk_cluster mode)
dataloader_params:
batch_size: 1
drop_last: True
num_workers: 0
pin_memory: False
55 changes: 55 additions & 0 deletions configs/dataset/graph/cocitation_pubmed_for_partitioning.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,55 @@
# Dataset loader config
loader:
_target_: topobench.data.loaders.PlanetoidDatasetLoader
parameters:
data_domain: graph
data_type: cocitation_on_disk
data_name: PubMed
data_dir: ${paths.data_dir}/${dataset.loader.parameters.data_domain}/${dataset.loader.parameters.data_type}

# Choose memory mode
memory_type: on_disk_cluster # "in_memory", "on_disk", or "on_disk_cluster"

# Global partition settings
cluster:
num_parts: 64 # number of clusters for single global partition
recursive: false # forwarded to PyG ClusterData
keep_inter_cluster_edges: false # standard Cluster-GCN style
sparse_format: csr # required by our block-stream loader

# Streaming / loader settings for block-wise training
stream:
q: 4 # clusters per batch (Cluster-GCN "bsize")
num_workers: 0 # increase if needed
pin_memory: false # true if you want faster H2D
with_edge_attr: false # WebKB has no edge_attr
precompute_split_parts: true # write parts_with_{train,val,test}.npy

# How to store features on disk
dtype_policy: float32 # current ClusterOnDisk uses float32

# Dataset parameters (unchanged)
parameters:
num_features: 500
num_classes: 3
num_nodes: 19717
task: classification
loss_type: cross_entropy
monitor_metric: accuracy
task_level: node

# Splits (unchanged; used by pack_global_partition)
split_params:
learning_setting: transductive # transductive # inductive
data_split_dir: ${paths.data_dir}/data_splits/${dataset.loader.parameters.data_name}
data_seed: 0
split_type: random
k: 10
train_prop: 0.5

# Dataloader parameters for legacy paths (kept; ignored in on_disk_cluster mode)
dataloader_params:
batch_size: 1
drop_last: True
num_workers: 0
pin_memory: False
55 changes: 55 additions & 0 deletions configs/dataset/graph/cocitation_pubmed_on_disk.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,55 @@
# Dataset loader config
loader:
_target_: topobench.data.loaders.OnDiskDatasetLoader
parameters:
data_domain: graph
data_type: cocitation_on_disk
data_name: PubMed_on_disk
data_dir: ${paths.data_dir}/${dataset.loader.parameters.data_domain}/${dataset.loader.parameters.data_type}

# Choose memory mode
memory_type: on_disk # "in_memory", "on_disk", or "on_disk_cluster"

# Global partition setting
cluster:
num_parts: 64 # number of clusters for single global partition
recursive: false # forwarded to PyG ClusterData
keep_inter_cluster_edges: false # standard Cluster-GCN style
sparse_format: csr # required by our block-stream loader

# Streaming / loader settings for block-wise training
stream:
q: 4 # clusters per batch (Cluster-GCN "bsize")
num_workers: 0 # increase if needed
pin_memory: false # true if you want faster H2D
with_edge_attr: false # WebKB has no edge_attr
precompute_split_parts: true # write parts_with_{train,val,test}.npy

# How to store features on disk
dtype_policy: float32 # current ClusterOnDisk uses float32

# Dataset parameters (unchanged)
parameters:
num_features: 500
num_classes: 3
num_nodes: 19717
task: classification
loss_type: cross_entropy
monitor_metric: accuracy
task_level: node

# Splits (unchanged; used by pack_global_partition)
split_params:
learning_setting: transductive # transductive # inductive
data_split_dir: ${paths.data_dir}/data_splits/${dataset.loader.parameters.data_name}
data_seed: 0
split_type: random
k: 10
train_prop: 0.5

# Dataloader parameters for legacy paths (kept; ignored in on_disk_cluster mode)
dataloader_params:
batch_size: 1
drop_last: True
num_workers: 0
pin_memory: False
Loading