Fix URL for summarization dataset (stanford-crfm#1935)

yifanmai · web-flow · commit 1aa727b67508 · 2023-10-23T10:20:13.000-07:00
diff --git a/src/helm/benchmark/scenarios/summarization_scenario.py b/src/helm/benchmark/scenarios/summarization_scenario.py
@@ -119,7 +119,7 @@ def _load_dataset(self, dataset_name: str, output_path: str):
             summary_key = "summary"
         elif dataset_name == "cnn-dm":
             url = (
-                "https://storage.cloud.google.com/crfm-helm-public/source_datasets/"
+                "https://storage.googleapis.com/crfm-helm-public/source_datasets/"
                 "scenarios/summarization_scenario/cnndm.pk"
             )
             dataset = self._download_dataset(url, "cnndm", output_path)

Original file line number	Diff line number	Diff line change
`@@ -119,7 +119,7 @@ def _load_dataset(self, dataset_name: str, output_path: str):`
`119`	`119`	`summary_key = "summary"`
`120`	`120`	`elif dataset_name == "cnn-dm":`
`121`	`121`	`url = (`
`122`		`- "https://storage.cloud.google.com/crfm-helm-public/source_datasets/"`
	`122`	`+ "https://storage.googleapis.com/crfm-helm-public/source_datasets/"`
`123`	`123`	`"scenarios/summarization_scenario/cnndm.pk"`
`124`	`124`	`)`
`125`	`125`	`dataset = self._download_dataset(url, "cnndm", output_path)`