Fix the typo of command running papers100M (#1121)

Fix the typo of command running papers100M and test it with the docker for dataset downloading, preprocessing, and evaluation. --------- Co-authored-by: Ubuntu <[email protected]>
awslabs · Jan 1, 2025 · 70f70b7 · 70f70b7
1 parent ab784a6
commit 70f70b7
Show file tree

Hide file tree

Showing 4 changed files with 6 additions and 3 deletions.
diff --git a/.gitignore b/.gitignore
@@ -27,4 +27,6 @@ docs/source/generated
 # IDEs and Python venv
 .idea
 .venv
-.vscode
+.vscode
+
+dataset/*
diff --git a/graphstorm_job.sh b/graphstorm_job.sh
diff --git a/python/graphstorm/data/ogbn_datasets.py b/python/graphstorm/data/ogbn_datasets.py
@@ -86,7 +86,7 @@ def __init__(self, raw_dir, dataset, edge_pct=1,
             self._num_classes = 47
         elif dataset == "ogbn-arxiv":
             self._num_classes = 40
-        elif dataset == "ogbn-papers100M":
+        elif dataset == "ogbn-papers100M" or dataset == "ogbn-papers100m":
             self._num_classes = 172
         self.retain_original_features = retain_original_features
         self.lm_model_name=lm_model_name

diff --git a/tools/partition_graph.py b/tools/partition_graph.py
@@ -121,7 +121,8 @@
                                      retain_original_features=args.retain_original_features,
                                      max_sequence_length=args.max_seq_length,
                                      lm_model_name=args.lm_model_name)
-    elif args.dataset == 'ogbn-papers100m':
+    elif args.dataset == 'ogbn-papers100M' or args.dataset == 'ogbn-papers100m':
+        args.dataset = 'ogbn-papers100M'
         dataset = OGBTextFeatDataset(args.filepath, dataset=args.dataset,
                                      retain_original_features=args.retain_original_features,
                                      max_sequence_length=args.max_seq_length,
-Original file line number
+Diff line change
@@ Expand Up / @@ -27,4 +27,6 @@ docs/source/generated @@
     # IDEs and Python venv
     .idea
     .venv
-    .vscode
+    .vscode
+    dataset/*