Spaces:
Runtime error
Runtime error
Update Space (evaluate main: c447fc8e)
Browse files- mauve.py +38 -54
- requirements.txt +1 -1
mauve.py
CHANGED
@@ -14,9 +14,6 @@
|
|
14 |
# limitations under the License.
|
15 |
""" MAUVE metric from https://github.com/krishnap25/mauve. """
|
16 |
|
17 |
-
from dataclasses import dataclass
|
18 |
-
from typing import List, Optional, Union
|
19 |
-
|
20 |
import datasets
|
21 |
import faiss # Here to have a nice missing dependency error message early on
|
22 |
import numpy # Here to have a nice missing dependency error message early on
|
@@ -88,47 +85,14 @@ Examples:
|
|
88 |
"""
|
89 |
|
90 |
|
91 |
-
@dataclass
|
92 |
-
class MauveConfig(evaluate.info.Config):
|
93 |
-
|
94 |
-
name: str = "default"
|
95 |
-
|
96 |
-
pos_label: Union[str, int] = 1
|
97 |
-
average: str = "binary"
|
98 |
-
labels: Optional[List[str]] = None
|
99 |
-
sample_weight: Optional[List[float]] = None
|
100 |
-
|
101 |
-
p_features: Optional[List] = None
|
102 |
-
q_features: Optional[List] = None
|
103 |
-
p_tokens: Optional[List] = None
|
104 |
-
q_tokens: Optional[List] = None
|
105 |
-
num_buckets: str = "auto"
|
106 |
-
pca_max_data: int = -1
|
107 |
-
kmeans_explained_var: float = 0.9
|
108 |
-
kmeans_num_redo: int = 5
|
109 |
-
kmeans_max_iter: int = 500
|
110 |
-
featurize_model_name: str = "gpt2-large"
|
111 |
-
device_id: int = (-1,)
|
112 |
-
max_text_length: int = 1024
|
113 |
-
divergence_curve_discretization_size: int = 25
|
114 |
-
mauve_scaling_factor: int = 5
|
115 |
-
verbose: bool = True
|
116 |
-
seed: int = 25
|
117 |
-
|
118 |
-
|
119 |
@evaluate.utils.file_utils.add_start_docstrings(_DESCRIPTION, _KWARGS_DESCRIPTION)
|
120 |
class Mauve(evaluate.Metric):
|
121 |
-
|
122 |
-
CONFIG_CLASS = MauveConfig
|
123 |
-
ALLOWED_CONFIG_NAMES = ["default"]
|
124 |
-
|
125 |
-
def _info(self, config):
|
126 |
return evaluate.MetricInfo(
|
127 |
description=_DESCRIPTION,
|
128 |
citation=_CITATION,
|
129 |
homepage="https://github.com/krishnap25/mauve",
|
130 |
inputs_description=_KWARGS_DESCRIPTION,
|
131 |
-
config=config,
|
132 |
features=datasets.Features(
|
133 |
{
|
134 |
"predictions": datasets.Value("string", id="sequence"),
|
@@ -142,25 +106,45 @@ class Mauve(evaluate.Metric):
|
|
142 |
],
|
143 |
)
|
144 |
|
145 |
-
def _compute(
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
146 |
out = compute_mauve(
|
147 |
p_text=predictions,
|
148 |
q_text=references,
|
149 |
-
p_features=
|
150 |
-
q_features=
|
151 |
-
p_tokens=
|
152 |
-
q_tokens=
|
153 |
-
num_buckets=
|
154 |
-
pca_max_data=
|
155 |
-
kmeans_explained_var=
|
156 |
-
kmeans_num_redo=
|
157 |
-
kmeans_max_iter=
|
158 |
-
featurize_model_name=
|
159 |
-
device_id=
|
160 |
-
max_text_length=
|
161 |
-
divergence_curve_discretization_size=
|
162 |
-
mauve_scaling_factor=
|
163 |
-
verbose=
|
164 |
-
seed=
|
165 |
)
|
166 |
return out
|
|
|
14 |
# limitations under the License.
|
15 |
""" MAUVE metric from https://github.com/krishnap25/mauve. """
|
16 |
|
|
|
|
|
|
|
17 |
import datasets
|
18 |
import faiss # Here to have a nice missing dependency error message early on
|
19 |
import numpy # Here to have a nice missing dependency error message early on
|
|
|
85 |
"""
|
86 |
|
87 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
88 |
@evaluate.utils.file_utils.add_start_docstrings(_DESCRIPTION, _KWARGS_DESCRIPTION)
|
89 |
class Mauve(evaluate.Metric):
|
90 |
+
def _info(self):
|
|
|
|
|
|
|
|
|
91 |
return evaluate.MetricInfo(
|
92 |
description=_DESCRIPTION,
|
93 |
citation=_CITATION,
|
94 |
homepage="https://github.com/krishnap25/mauve",
|
95 |
inputs_description=_KWARGS_DESCRIPTION,
|
|
|
96 |
features=datasets.Features(
|
97 |
{
|
98 |
"predictions": datasets.Value("string", id="sequence"),
|
|
|
106 |
],
|
107 |
)
|
108 |
|
109 |
+
def _compute(
|
110 |
+
self,
|
111 |
+
predictions,
|
112 |
+
references,
|
113 |
+
p_features=None,
|
114 |
+
q_features=None,
|
115 |
+
p_tokens=None,
|
116 |
+
q_tokens=None,
|
117 |
+
num_buckets="auto",
|
118 |
+
pca_max_data=-1,
|
119 |
+
kmeans_explained_var=0.9,
|
120 |
+
kmeans_num_redo=5,
|
121 |
+
kmeans_max_iter=500,
|
122 |
+
featurize_model_name="gpt2-large",
|
123 |
+
device_id=-1,
|
124 |
+
max_text_length=1024,
|
125 |
+
divergence_curve_discretization_size=25,
|
126 |
+
mauve_scaling_factor=5,
|
127 |
+
verbose=True,
|
128 |
+
seed=25,
|
129 |
+
):
|
130 |
out = compute_mauve(
|
131 |
p_text=predictions,
|
132 |
q_text=references,
|
133 |
+
p_features=p_features,
|
134 |
+
q_features=q_features,
|
135 |
+
p_tokens=p_tokens,
|
136 |
+
q_tokens=q_tokens,
|
137 |
+
num_buckets=num_buckets,
|
138 |
+
pca_max_data=pca_max_data,
|
139 |
+
kmeans_explained_var=kmeans_explained_var,
|
140 |
+
kmeans_num_redo=kmeans_num_redo,
|
141 |
+
kmeans_max_iter=kmeans_max_iter,
|
142 |
+
featurize_model_name=featurize_model_name,
|
143 |
+
device_id=device_id,
|
144 |
+
max_text_length=max_text_length,
|
145 |
+
divergence_curve_discretization_size=divergence_curve_discretization_size,
|
146 |
+
mauve_scaling_factor=mauve_scaling_factor,
|
147 |
+
verbose=verbose,
|
148 |
+
seed=seed,
|
149 |
)
|
150 |
return out
|
requirements.txt
CHANGED
@@ -1,4 +1,4 @@
|
|
1 |
-
git+https://github.com/huggingface/evaluate@
|
2 |
faiss-cpu
|
3 |
sklearn
|
4 |
mauve-text
|
|
|
1 |
+
git+https://github.com/huggingface/evaluate@c447fc8eda9c62af501bfdc6988919571050d950
|
2 |
faiss-cpu
|
3 |
sklearn
|
4 |
mauve-text
|