0-shot description

README.md

@@ -1,8 +1,8 @@
 ---
-title: Demo Leaderboard
-emoji: 🥇
-colorFrom: green
-colorTo: indigo
+title: Open PL LLM Leaderboard
+emoji: 🏆🇵🇱
+colorFrom: gray
+colorTo: red
 sdk: gradio
 sdk_version: 4.4.0
 app_file: app.py

src/about.py

@@ -37,7 +37,7 @@ NUM_FEWSHOT = 0 # Change with your few shot
 
 
 # Your leaderboard name
-TITLE = """<h1 align="center" id="space-title">Open PL LLM Leaderboard</h1>"""
+TITLE = """<h1 align="center" id="space-title">Open PL LLM Leaderboard (0-shot)</h1>"""
 
 # What does your leaderboard evaluate?
 INTRODUCTION_TEXT = """

src/leaderboard/read_evals.py

@@ -12,6 +12,7 @@ from src.display.formatting import make_clickable_model
 from src.display.utils import AutoEvalColumn, ModelType, Tasks, Precision, WeightType
 from src.submission.check_validity import is_model_on_hub
 
+NUM_FEWSHOT = 0
 
 @dataclass
 class EvalResult:
@@ -73,7 +74,7 @@ class EvalResult:
             task = task.value
 
             # We average all scores of a given metric (not all metrics are present in all files)
-            accs = np.array([v.get(task.metric, None) for k, v in data["results"].items() if task.benchmark == k and n_shot.get(k, -1) == 0])
+            accs = np.array([v.get(task.metric, None) for k, v in data["results"].items() if task.benchmark == k and n_shot.get(k, -1) == NUM_FEWSHOT])
             if accs.size == 0 or any([acc is None for acc in accs]):
                 continue