Upload folder using huggingface_hub

inference.py

@@ -68,13 +68,13 @@ def emit_step(step_no: int, action_text: str, reward: float, done: bool, error:
     )
 
 
-def emit_end(success: bool, steps: int, score: float, rewards: List[float]) -> None:
-    reward_text = ",".join(f"{reward:.2f}" for reward in rewards)
-    print(
-        f"[END] success={str(success).lower()} steps={steps} "
-        f"score={score:.2f} rewards={reward_text}",
-        flush=True,
-    )
+def emit_end(success: bool, steps: int, score: float, rewards: List[float]) -> None:
+    reward_text = ",".join(f"{reward:.2f}" for reward in rewards)
+    print(
+        f"[END] success={str(success).lower()} steps={steps} "
+        f"score={score:.6f} rewards={reward_text}",
+        flush=True,
+    )
 
 
 def choose_tasks(selection: str) -> Iterable[str]:
@@ -142,11 +142,9 @@ def compact_action(action: PharmaAction) -> str:
     return label
 
 
-def final_score(rewards: List[float]) -> float:
-    if not rewards:
-        return 0.0
-    score = sum(rewards) / len(rewards)
-    return min(max(round(score, 4), 0.0), 1.0)
+def final_score(rewards: List[float]) -> float:
+    score = sum(rewards) / len(rewards) if rewards else 0.0
+    return max(1e-6, min(score, 1 - 1e-6))
 
 
 def run_one_task(llm: OpenAI, task_name: str) -> None: