Fix AIME25 naming (#270)

* Fix AIME25 naming * Fix
huggingface · Feb 10, 2025 · 37d9887 · 37d9887
1 parent baec330
commit 37d9887
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 4 deletions.
diff --git a/src/open_r1/evaluate.py b/src/open_r1/evaluate.py
@@ -106,8 +106,9 @@ def gpqa_prompt_fn(line, task_name: str = None):
     metric=[expr_gold_metric],
     version=1,
 )
-aime25 = LightevalTaskConfig(
-    name="aime25",
+# Part I from AIME 2025 exam: https://artofproblemsolving.com/wiki/index.php/2025_AIME_I?srsltid=AfmBOoof5gaaqlt3-l6LH7Tt6qmJZtl_2PQEDYlLFlMqhq9dLL8FMCRR
+aime25_part1 = LightevalTaskConfig(
+    name="aime25:part1",
     suite=["custom"],
     prompt_function=aime_prompt_fn,
     hf_repo="open-r1/aime_2025_1",
@@ -155,7 +156,7 @@ def gpqa_prompt_fn(line, task_name: str = None):
 # Add tasks to the table
 TASKS_TABLE = []
 TASKS_TABLE.append(aime24)
-TASKS_TABLE.append(aime25)
+TASKS_TABLE.append(aime25_part1)
 TASKS_TABLE.append(math_500)
 TASKS_TABLE.append(gpqa_diamond)
 

diff --git a/src/open_r1/utils/evaluation.py b/src/open_r1/utils/evaluation.py
@@ -48,7 +48,7 @@ def register_lighteval_task(
 
 register_lighteval_task(LIGHTEVAL_TASKS, "custom", "math_500", "math_500", 0)
 register_lighteval_task(LIGHTEVAL_TASKS, "custom", "aime24", "aime24", 0)
-register_lighteval_task(LIGHTEVAL_TASKS, "custom", "aime25", "aime25", 0)
+register_lighteval_task(LIGHTEVAL_TASKS, "custom", "aime25_part1", "aime25:part1", 0)
 register_lighteval_task(LIGHTEVAL_TASKS, "custom", "gpqa", "gpqa:diamond", 0)