refactor: linting

IAAR-Shanghai · Aug 20, 2024 · 23203e8 · 23203e8
1 parent 6b7d43f
commit 23203e8
Show file tree

Hide file tree

Showing 4 changed files with 5 additions and 6 deletions.
diff --git a/eval/benchs/halueval/eval_halueval_dialog.py b/eval/benchs/halueval/eval_halueval_dialog.py
@@ -51,7 +51,7 @@ def scoring(self, data_point: dict) -> dict:
         )
         response = self.model.safe_request(query)
 
-        answer = response.strip().split()  
+        answer = response.strip().split()
         # Extract the first word, such as "Yes", "No", "#Yes", "No."
         # Note: "".strip() returns [] instead of [""]
         answer = answer[0] if answer else ""

diff --git a/eval/benchs/halueval/eval_halueval_qa.py b/eval/benchs/halueval/eval_halueval_qa.py
@@ -56,14 +56,13 @@ def scoring(self, data_point: dict) -> dict:
         )
         response = self.model.safe_request(query)
 
-        answer = response.strip().split()  
+        answer = response.strip().split()
         # Extract the first word, such as "Yes", "No", "#Yes", "No."
         # Note: "".strip() returns [] instead of [""]
         answer = answer[0] if answer else ""
         # Remove the leading "#", ".", ","
         answer = answer.strip("#").strip(".").strip(",")
 
-
         return {
             "metrics": {
                 "correct": ground_truth.lower() == answer.lower(),

diff --git a/eval/benchs/halueval/eval_halueval_summa.py b/eval/benchs/halueval/eval_halueval_summa.py
@@ -42,7 +42,7 @@ def scoring(self, data_point: dict) -> dict:
         )
         response = self.model.safe_request(query)
 
-        answer = response.strip().split()  
+        answer = response.strip().split()
         # Extract the first word, such as "Yes", "No", "#Yes", "No."
         # Note: "".strip() returns [] instead of [""]
         answer = answer[0] if answer else ""

diff --git a/eval/benchs/uhgeval/eval_sele.py b/eval/benchs/uhgeval/eval_sele.py
@@ -41,8 +41,8 @@ def which_is_true(self, contn1: str, contn2: str, obj: dict) -> tuple[int, str]:
         """Given two continuations, determine which one is more accurate.
 
         Returns:
-            tuple: (answer, response). `answer` is 1 if the first continuation is more 
-            accurate, 2 if the second one is more accurate, and -1 if error. `response` 
+            tuple: (answer, response). `answer` is 1 if the first continuation is more
+            accurate, 2 if the second one is more accurate, and -1 if error. `response`
             is the model's response.
         """
         query = PROMPT_TEMPLATE.format(