murinj
commited on
Commit
·
d275ac7
1
Parent(s):
d27d43d
init
Browse files- .gitignore +1 -0
- HTER.py +98 -0
- README.md +84 -6
- app.py +6 -0
- requirements.txt +0 -0
.gitignore
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
/.idea
|
HTER.py
ADDED
@@ -0,0 +1,98 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
|
2 |
+
#
|
3 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
4 |
+
# you may not use this file except in compliance with the License.
|
5 |
+
# You may obtain a copy of the License at
|
6 |
+
#
|
7 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
8 |
+
#
|
9 |
+
# Unless required by applicable law or agreed to in writing, software
|
10 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
11 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
12 |
+
# See the License for the specific language governing permissions and
|
13 |
+
# limitations under the License.
|
14 |
+
"""HTER metric."""
|
15 |
+
|
16 |
+
import datasets
|
17 |
+
|
18 |
+
|
19 |
+
import evaluate
|
20 |
+
|
21 |
+
|
22 |
+
_DESCRIPTION = """
|
23 |
+
HTER (Half Total Error Rate) is a metric that combines the False Accept Rate (FAR) and False Reject Rate (FRR) to provide a comprehensive evaluation of a system's performance. It can be computed with:
|
24 |
+
HTER = (FAR + FRR) / 2
|
25 |
+
Where:
|
26 |
+
FAR (False Accept Rate) = FP / (FP + TN)
|
27 |
+
FRR (False Reject Rate) = FN / (FN + TP)
|
28 |
+
TP: True positive
|
29 |
+
TN: True negative
|
30 |
+
FP: False positive
|
31 |
+
FN: False negative
|
32 |
+
"""
|
33 |
+
|
34 |
+
|
35 |
+
_KWARGS_DESCRIPTION = """
|
36 |
+
Args:
|
37 |
+
predictions (`list` of `int`): Predicted labels.
|
38 |
+
references (`list` of `int`): Ground truth labels.
|
39 |
+
|
40 |
+
Returns:
|
41 |
+
HTER (`float` or `int`): HTER score. Minimum possible value is 0. Maximum possible value is 1.0, or the number of examples input.
|
42 |
+
|
43 |
+
Examples:
|
44 |
+
|
45 |
+
Example 1-A simple example
|
46 |
+
>>> hter_metric = evaluate.load("hter")
|
47 |
+
>>> results = hter_metric.compute(references=[0, 0], predictions=[0, 1])
|
48 |
+
>>> print(results)
|
49 |
+
{'HTER': 0.5}
|
50 |
+
|
51 |
+
"""
|
52 |
+
|
53 |
+
|
54 |
+
_CITATION = """
|
55 |
+
}
|
56 |
+
"""
|
57 |
+
|
58 |
+
|
59 |
+
@evaluate.utils.file_utils.add_start_docstrings(_DESCRIPTION, _KWARGS_DESCRIPTION)
|
60 |
+
class Accuracy(evaluate.Metric):
|
61 |
+
def _info(self):
|
62 |
+
return evaluate.MetricInfo(
|
63 |
+
description=_DESCRIPTION,
|
64 |
+
citation=_CITATION,
|
65 |
+
inputs_description=_KWARGS_DESCRIPTION,
|
66 |
+
features=datasets.Features(
|
67 |
+
{
|
68 |
+
"predictions": datasets.Value("int32"),
|
69 |
+
"references": datasets.Value("int32"),
|
70 |
+
}
|
71 |
+
),
|
72 |
+
)
|
73 |
+
|
74 |
+
def _compute(self, predictions, references):
|
75 |
+
TP = 0 # True Positive
|
76 |
+
TN = 0 # True Negative
|
77 |
+
FP = 0 # False Positive
|
78 |
+
FN = 0 # False Negative
|
79 |
+
|
80 |
+
for pred, ref in zip(predictions, references):
|
81 |
+
if pred == 1 and ref == 1:
|
82 |
+
TP += 1
|
83 |
+
elif pred == 0 and ref == 0:
|
84 |
+
TN += 1
|
85 |
+
elif pred == 1 and ref == 0:
|
86 |
+
FP += 1
|
87 |
+
elif pred == 0 and ref == 1:
|
88 |
+
FN += 1
|
89 |
+
|
90 |
+
# 计算 FAR 和 FRR
|
91 |
+
FAR = FP / (FP + TN) if (FP + TN) > 0 else 0
|
92 |
+
FRR = FN / (FN + TP) if (FN + TP) > 0 else 0
|
93 |
+
|
94 |
+
# 计算 HTER
|
95 |
+
HTER = (FAR + FRR) / 2.0
|
96 |
+
return {
|
97 |
+
"HTER": HTER
|
98 |
+
}
|
README.md
CHANGED
@@ -1,12 +1,90 @@
|
|
1 |
---
|
2 |
-
title:
|
3 |
-
emoji:
|
4 |
-
colorFrom:
|
5 |
-
colorTo:
|
6 |
sdk: gradio
|
7 |
-
sdk_version:
|
8 |
app_file: app.py
|
9 |
pinned: false
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
---
|
11 |
|
12 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
+
title: HTER
|
3 |
+
emoji: 🤗
|
4 |
+
colorFrom: blue
|
5 |
+
colorTo: red
|
6 |
sdk: gradio
|
7 |
+
sdk_version: 3.19.1
|
8 |
app_file: app.py
|
9 |
pinned: false
|
10 |
+
tags:
|
11 |
+
- evaluate
|
12 |
+
- metric
|
13 |
+
description: >-
|
14 |
+
HTER (Half Total Error Rate) is a metric that combines the False Accept Rate (FAR) and False Reject Rate (FRR) to provide a comprehensive evaluation of a system's performance. It can be computed with:
|
15 |
+
HTER = (FAR + FRR) / 2
|
16 |
+
Where:
|
17 |
+
FAR (False Accept Rate) = FP / (FP + TN)
|
18 |
+
FRR (False Reject Rate) = FN / (FN + TP)
|
19 |
+
TP: True positive
|
20 |
+
TN: True negative
|
21 |
+
FP: False positive
|
22 |
+
FN: False negative
|
23 |
---
|
24 |
|
25 |
+
# Metric Card for HTER
|
26 |
+
|
27 |
+
|
28 |
+
## Metric Description
|
29 |
+
|
30 |
+
HTER (Half Total Error Rate) is a metric that combines the False Accept Rate (FAR) and False Reject Rate (FRR) to provide a comprehensive evaluation of a system's performance. It can be computed with:
|
31 |
+
|
32 |
+
HTER = (FAR + FRR) / 2
|
33 |
+
|
34 |
+
Where:
|
35 |
+
|
36 |
+
FAR (False Accept Rate) = FP / (FP + TN)
|
37 |
+
|
38 |
+
FRR (False Reject Rate) = FN / (FN + TP)
|
39 |
+
|
40 |
+
TP: True positive
|
41 |
+
|
42 |
+
TN: True negative
|
43 |
+
|
44 |
+
FP: False positive
|
45 |
+
|
46 |
+
FN: False negative
|
47 |
+
|
48 |
+
|
49 |
+
|
50 |
+
## How to Use
|
51 |
+
|
52 |
+
At minimum, this metric requires predictions and references as inputs.
|
53 |
+
|
54 |
+
```python
|
55 |
+
>>> hter_metric = evaluate.load("hter")
|
56 |
+
>>> results = hter_metric.compute(references=[0, 0], predictions=[0, 1])
|
57 |
+
>>> print(results)
|
58 |
+
{'HTER': 0.5}
|
59 |
+
```
|
60 |
+
|
61 |
+
|
62 |
+
### Inputs
|
63 |
+
- **predictions** (`list` of `int`): Predicted labels.
|
64 |
+
- **references** (`list` of `int`): Ground truth labels.
|
65 |
+
|
66 |
+
[//]: # (- **normalize** (`boolean`): If set to False, returns the number of correctly classified samples. Otherwise, returns the fraction of correctly classified samples. Defaults to True.)
|
67 |
+
|
68 |
+
[//]: # (- **sample_weight** (`list` of `float`): Sample weights Defaults to None.)
|
69 |
+
|
70 |
+
|
71 |
+
### Output Values
|
72 |
+
- **HTER**(`float` or `int`): HTER score. Minimum possible value is 0. Maximum possible value is 1.0, or the number of examples input.
|
73 |
+
|
74 |
+
Output Example(s):
|
75 |
+
```python
|
76 |
+
{'HTER': 1.0}
|
77 |
+
```
|
78 |
+
|
79 |
+
This metric outputs a dictionary, containing the HTER score.
|
80 |
+
|
81 |
+
|
82 |
+
[//]: # (## Citation(s))
|
83 |
+
|
84 |
+
[//]: # (```bibtex)
|
85 |
+
|
86 |
+
[//]: # ()
|
87 |
+
[//]: # (```)
|
88 |
+
|
89 |
+
|
90 |
+
## Further References
|
app.py
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import evaluate
|
2 |
+
from evaluate.utils import launch_gradio_widget
|
3 |
+
|
4 |
+
|
5 |
+
module = evaluate.load("hter")
|
6 |
+
launch_gradio_widget(module)
|
requirements.txt
ADDED
File without changes
|