Datasets:
Tasks:
Text Classification
Sub-tasks:
multi-class-classification
Languages:
English
Size:
100K<n<1M
ArXiv:
Tags:
relation extraction
License:
updated docu
Browse files
tacred.py
CHANGED
@@ -15,7 +15,6 @@
|
|
15 |
|
16 |
"""The TACRED dataset for English Relation Classification"""
|
17 |
|
18 |
-
|
19 |
import json
|
20 |
import os
|
21 |
|
@@ -67,6 +66,12 @@ TACRED is a large-scale relation extraction dataset with 106,264 examples built
|
|
67 |
Note: There is currently a label-corrected version of the TACRED dataset, which you should consider using instead of
|
68 |
the original version released in 2017. For more details on this new version, see the TACRED Revisited paper
|
69 |
published at ACL 2020.
|
|
|
|
|
|
|
|
|
|
|
|
|
70 |
"""
|
71 |
|
72 |
_HOMEPAGE = "https://nlp.stanford.edu/projects/tacred/"
|
@@ -82,6 +87,8 @@ _PATCH_URLs = {
|
|
82 |
"test": "https://raw.githubusercontent.com/DFKI-NLP/tacrev/master/patch/test_patch.json",
|
83 |
}
|
84 |
|
|
|
|
|
85 |
_CLASS_LABELS = [
|
86 |
"no_relation",
|
87 |
"org:alternate_names",
|
@@ -146,11 +153,11 @@ class Tacred(datasets.GeneratorBasedBuilder):
|
|
146 |
|
147 |
BUILDER_CONFIGS = [
|
148 |
datasets.BuilderConfig(
|
149 |
-
name="original", version=
|
150 |
),
|
151 |
datasets.BuilderConfig(
|
152 |
name="revised",
|
153 |
-
version=
|
154 |
description="The revised TACRED (corrected labels in dev and test split).",
|
155 |
),
|
156 |
]
|
|
|
15 |
|
16 |
"""The TACRED dataset for English Relation Classification"""
|
17 |
|
|
|
18 |
import json
|
19 |
import os
|
20 |
|
|
|
66 |
Note: There is currently a label-corrected version of the TACRED dataset, which you should consider using instead of
|
67 |
the original version released in 2017. For more details on this new version, see the TACRED Revisited paper
|
68 |
published at ACL 2020.
|
69 |
+
|
70 |
+
Note 2: This Datasetreader changes the offsets of the following fields, to conform with standard Python usage (see
|
71 |
+
#_generate_examples()):
|
72 |
+
- subj_end to subj_end + 1 (make end offset exclusive)
|
73 |
+
- oj_end to oj_end + 1 (make end offset exclusive)
|
74 |
+
- stanford_head to stanford_head - 1 (make head offsets 0-based)
|
75 |
"""
|
76 |
|
77 |
_HOMEPAGE = "https://nlp.stanford.edu/projects/tacred/"
|
|
|
87 |
"test": "https://raw.githubusercontent.com/DFKI-NLP/tacrev/master/patch/test_patch.json",
|
88 |
}
|
89 |
|
90 |
+
_VERSION = datasets.Version("1.0.0")
|
91 |
+
|
92 |
_CLASS_LABELS = [
|
93 |
"no_relation",
|
94 |
"org:alternate_names",
|
|
|
153 |
|
154 |
BUILDER_CONFIGS = [
|
155 |
datasets.BuilderConfig(
|
156 |
+
name="original", version=_VERSION, description="The original TACRED."
|
157 |
),
|
158 |
datasets.BuilderConfig(
|
159 |
name="revised",
|
160 |
+
version=_VERSION,
|
161 |
description="The revised TACRED (corrected labels in dev and test split).",
|
162 |
),
|
163 |
]
|