Spaces:
Running
on
CPU Upgrade
Running
on
CPU Upgrade
pminervini
commited on
Commit
•
b25a00b
1
Parent(s):
d265631
update
Browse files
src/backend/tasks/halueval/halueval_dialogue.yaml
CHANGED
@@ -7,8 +7,8 @@ validation_split: data
|
|
7 |
test_split: data
|
8 |
num_fewshot: 0
|
9 |
doc_to_text: !function utils.doc_to_text_dialogue
|
10 |
-
doc_to_target: !function utils.
|
11 |
-
process_results: !function utils.
|
12 |
metric_list:
|
13 |
- metric: em
|
14 |
aggregation: mean
|
|
|
7 |
test_split: data
|
8 |
num_fewshot: 0
|
9 |
doc_to_text: !function utils.doc_to_text_dialogue
|
10 |
+
doc_to_target: !function utils.doc_to_target
|
11 |
+
process_results: !function utils.process_results
|
12 |
metric_list:
|
13 |
- metric: em
|
14 |
aggregation: mean
|
src/backend/tasks/halueval/halueval_qa.yaml
CHANGED
@@ -7,8 +7,8 @@ validation_split: data
|
|
7 |
test_split: data
|
8 |
num_fewshot: 0
|
9 |
doc_to_text: !function utils.doc_to_text_qa
|
10 |
-
doc_to_target: !function utils.
|
11 |
-
process_results: !function utils.
|
12 |
metric_list:
|
13 |
- metric: em
|
14 |
aggregation: mean
|
|
|
7 |
test_split: data
|
8 |
num_fewshot: 0
|
9 |
doc_to_text: !function utils.doc_to_text_qa
|
10 |
+
doc_to_target: !function utils.doc_to_target
|
11 |
+
process_results: !function utils.process_results
|
12 |
metric_list:
|
13 |
- metric: em
|
14 |
aggregation: mean
|
src/backend/tasks/halueval/halueval_summarization.yaml
CHANGED
@@ -7,8 +7,8 @@ validation_split: data
|
|
7 |
test_split: data
|
8 |
num_fewshot: 0
|
9 |
doc_to_text: !function utils.doc_to_text_summarization
|
10 |
-
doc_to_target: !function utils.
|
11 |
-
process_results: !function utils.
|
12 |
metric_list:
|
13 |
- metric: em
|
14 |
aggregation: mean
|
|
|
7 |
test_split: data
|
8 |
num_fewshot: 0
|
9 |
doc_to_text: !function utils.doc_to_text_summarization
|
10 |
+
doc_to_target: !function utils.doc_to_target
|
11 |
+
process_results: !function utils.process_results
|
12 |
metric_list:
|
13 |
- metric: em
|
14 |
aggregation: mean
|
src/backend/tasks/halueval/utils.py
CHANGED
@@ -102,11 +102,11 @@ def doc_to_text_summarization(doc: dict[str, str]) -> str:
|
|
102 |
return doc_text
|
103 |
|
104 |
|
105 |
-
def
|
106 |
return doc['hallucination']
|
107 |
|
108 |
|
109 |
-
def
|
110 |
is_correct = True
|
111 |
|
112 |
if ("Yes" in prediction and "No" in prediction) or ("Yes" not in prediction and "No" not in prediction):
|
@@ -122,13 +122,15 @@ def compute_metrics_qa(gold_answer: str, prediction: str) -> dict[str, float]:
|
|
122 |
if is_correct:
|
123 |
res["em"] = 1.0 if is_exact else 0.0
|
124 |
|
|
|
|
|
125 |
return res
|
126 |
|
127 |
|
128 |
-
def
|
129 |
# results is e.g., ['Yes']
|
130 |
-
gold_list =
|
131 |
# gold_list is e.g., 'yes'
|
132 |
prediction = results[0].strip().split("\n")[0]
|
133 |
-
scores =
|
134 |
return scores
|
|
|
102 |
return doc_text
|
103 |
|
104 |
|
105 |
+
def doc_to_target(doc: dict[str, str]) -> str:
|
106 |
return doc['hallucination']
|
107 |
|
108 |
|
109 |
+
def compute_metrics(gold_answer: str, prediction: str) -> dict[str, float]:
|
110 |
is_correct = True
|
111 |
|
112 |
if ("Yes" in prediction and "No" in prediction) or ("Yes" not in prediction and "No" not in prediction):
|
|
|
122 |
if is_correct:
|
123 |
res["em"] = 1.0 if is_exact else 0.0
|
124 |
|
125 |
+
res["acc"] = 1.0 if (is_correct and is_exact) else 0.0
|
126 |
+
|
127 |
return res
|
128 |
|
129 |
|
130 |
+
def process_results(doc: dict[str, str], results: list[str]):
|
131 |
# results is e.g., ['Yes']
|
132 |
+
gold_list = doc_to_target(doc)
|
133 |
# gold_list is e.g., 'yes'
|
134 |
prediction = results[0].strip().split("\n")[0]
|
135 |
+
scores = compute_metrics(gold_list, prediction)
|
136 |
return scores
|