Spaces:
Running
Running
Update Space (evaluate main: c447fc8e)
Browse files- meteor.py +10 -33
- requirements.txt +1 -1
meteor.py
CHANGED
|
@@ -13,8 +13,6 @@
|
|
| 13 |
# limitations under the License.
|
| 14 |
""" METEOR metric. """
|
| 15 |
|
| 16 |
-
from dataclasses import dataclass
|
| 17 |
-
|
| 18 |
import datasets
|
| 19 |
import numpy as np
|
| 20 |
from datasets.config import importlib_metadata, version
|
|
@@ -84,28 +82,13 @@ Examples:
|
|
| 84 |
"""
|
| 85 |
|
| 86 |
|
| 87 |
-
@dataclass
|
| 88 |
-
class MeteorConfig(evaluate.info.Config):
|
| 89 |
-
|
| 90 |
-
name: str = "default"
|
| 91 |
-
|
| 92 |
-
alpha: float = 0.9
|
| 93 |
-
beta: float = 3.0
|
| 94 |
-
gamma: float = 0.5
|
| 95 |
-
|
| 96 |
-
|
| 97 |
@evaluate.utils.file_utils.add_start_docstrings(_DESCRIPTION, _KWARGS_DESCRIPTION)
|
| 98 |
class Meteor(evaluate.Metric):
|
| 99 |
-
|
| 100 |
-
CONFIG_CLASS = MeteorConfig
|
| 101 |
-
ALLOWED_CONFIG_NAMES = ["default", "multilabel"]
|
| 102 |
-
|
| 103 |
-
def _info(self, config):
|
| 104 |
return evaluate.MetricInfo(
|
| 105 |
description=_DESCRIPTION,
|
| 106 |
citation=_CITATION,
|
| 107 |
inputs_description=_KWARGS_DESCRIPTION,
|
| 108 |
-
config=config,
|
| 109 |
features=[
|
| 110 |
datasets.Features(
|
| 111 |
{
|
|
@@ -136,7 +119,7 @@ class Meteor(evaluate.Metric):
|
|
| 136 |
if NLTK_VERSION >= version.Version("3.6.6"):
|
| 137 |
nltk.download("omw-1.4")
|
| 138 |
|
| 139 |
-
def _compute(self, predictions, references):
|
| 140 |
multiple_refs = isinstance(references[0], list)
|
| 141 |
if NLTK_VERSION >= version.Version("3.6.5"):
|
| 142 |
# the version of METEOR in NLTK version 3.6.5 and earlier expect tokenized inputs
|
|
@@ -145,20 +128,16 @@ class Meteor(evaluate.Metric):
|
|
| 145 |
meteor_score.meteor_score(
|
| 146 |
[word_tokenize(ref) for ref in refs],
|
| 147 |
word_tokenize(pred),
|
| 148 |
-
alpha=
|
| 149 |
-
beta=
|
| 150 |
-
gamma=
|
| 151 |
)
|
| 152 |
for refs, pred in zip(references, predictions)
|
| 153 |
]
|
| 154 |
else:
|
| 155 |
scores = [
|
| 156 |
meteor_score.single_meteor_score(
|
| 157 |
-
word_tokenize(ref),
|
| 158 |
-
word_tokenize(pred),
|
| 159 |
-
alpha=self.config.alpha,
|
| 160 |
-
beta=self.config.beta,
|
| 161 |
-
gamma=self.config.gamma,
|
| 162 |
)
|
| 163 |
for ref, pred in zip(references, predictions)
|
| 164 |
]
|
|
@@ -168,17 +147,15 @@ class Meteor(evaluate.Metric):
|
|
| 168 |
meteor_score.meteor_score(
|
| 169 |
[[word_tokenize(ref) for ref in group] for group in references][0],
|
| 170 |
word_tokenize(pred),
|
| 171 |
-
alpha=
|
| 172 |
-
beta=
|
| 173 |
-
gamma=
|
| 174 |
)
|
| 175 |
for ref, pred in zip(references, predictions)
|
| 176 |
]
|
| 177 |
else:
|
| 178 |
scores = [
|
| 179 |
-
meteor_score.single_meteor_score(
|
| 180 |
-
ref, pred, alpha=self.config.alpha, beta=self.config.beta, gamma=self.config.gamma
|
| 181 |
-
)
|
| 182 |
for ref, pred in zip(references, predictions)
|
| 183 |
]
|
| 184 |
|
|
|
|
| 13 |
# limitations under the License.
|
| 14 |
""" METEOR metric. """
|
| 15 |
|
|
|
|
|
|
|
| 16 |
import datasets
|
| 17 |
import numpy as np
|
| 18 |
from datasets.config import importlib_metadata, version
|
|
|
|
| 82 |
"""
|
| 83 |
|
| 84 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 85 |
@evaluate.utils.file_utils.add_start_docstrings(_DESCRIPTION, _KWARGS_DESCRIPTION)
|
| 86 |
class Meteor(evaluate.Metric):
|
| 87 |
+
def _info(self):
|
|
|
|
|
|
|
|
|
|
|
|
|
| 88 |
return evaluate.MetricInfo(
|
| 89 |
description=_DESCRIPTION,
|
| 90 |
citation=_CITATION,
|
| 91 |
inputs_description=_KWARGS_DESCRIPTION,
|
|
|
|
| 92 |
features=[
|
| 93 |
datasets.Features(
|
| 94 |
{
|
|
|
|
| 119 |
if NLTK_VERSION >= version.Version("3.6.6"):
|
| 120 |
nltk.download("omw-1.4")
|
| 121 |
|
| 122 |
+
def _compute(self, predictions, references, alpha=0.9, beta=3, gamma=0.5):
|
| 123 |
multiple_refs = isinstance(references[0], list)
|
| 124 |
if NLTK_VERSION >= version.Version("3.6.5"):
|
| 125 |
# the version of METEOR in NLTK version 3.6.5 and earlier expect tokenized inputs
|
|
|
|
| 128 |
meteor_score.meteor_score(
|
| 129 |
[word_tokenize(ref) for ref in refs],
|
| 130 |
word_tokenize(pred),
|
| 131 |
+
alpha=alpha,
|
| 132 |
+
beta=beta,
|
| 133 |
+
gamma=gamma,
|
| 134 |
)
|
| 135 |
for refs, pred in zip(references, predictions)
|
| 136 |
]
|
| 137 |
else:
|
| 138 |
scores = [
|
| 139 |
meteor_score.single_meteor_score(
|
| 140 |
+
word_tokenize(ref), word_tokenize(pred), alpha=alpha, beta=beta, gamma=gamma
|
|
|
|
|
|
|
|
|
|
|
|
|
| 141 |
)
|
| 142 |
for ref, pred in zip(references, predictions)
|
| 143 |
]
|
|
|
|
| 147 |
meteor_score.meteor_score(
|
| 148 |
[[word_tokenize(ref) for ref in group] for group in references][0],
|
| 149 |
word_tokenize(pred),
|
| 150 |
+
alpha=alpha,
|
| 151 |
+
beta=beta,
|
| 152 |
+
gamma=gamma,
|
| 153 |
)
|
| 154 |
for ref, pred in zip(references, predictions)
|
| 155 |
]
|
| 156 |
else:
|
| 157 |
scores = [
|
| 158 |
+
meteor_score.single_meteor_score(ref, pred, alpha=alpha, beta=beta, gamma=gamma)
|
|
|
|
|
|
|
| 159 |
for ref, pred in zip(references, predictions)
|
| 160 |
]
|
| 161 |
|
requirements.txt
CHANGED
|
@@ -1,2 +1,2 @@
|
|
| 1 |
-
git+https://github.com/huggingface/evaluate@
|
| 2 |
nltk
|
|
|
|
| 1 |
+
git+https://github.com/huggingface/evaluate@c447fc8eda9c62af501bfdc6988919571050d950
|
| 2 |
nltk
|