Datasets:
Tasks:
Question Answering
Modalities:
Text
Formats:
parquet
Sub-tasks:
open-domain-qa
Languages:
English
Size:
10K - 100K
ArXiv:
License:
Commit
•
788845f
1
Parent(s):
72112ee
Add light data files
Browse files- README.md +10 -4
- dataset_infos.json +9 -36
- light/train-00000-of-00001.parquet +3 -0
- light/validation-00000-of-00001.parquet +3 -0
README.md
CHANGED
@@ -84,13 +84,13 @@ dataset_info:
|
|
84 |
sequence: string
|
85 |
splits:
|
86 |
- name: train
|
87 |
-
num_bytes:
|
88 |
num_examples: 10036
|
89 |
- name: validation
|
90 |
-
num_bytes:
|
91 |
num_examples: 2002
|
92 |
-
download_size:
|
93 |
-
dataset_size:
|
94 |
configs:
|
95 |
- config_name: full
|
96 |
data_files:
|
@@ -99,6 +99,12 @@ configs:
|
|
99 |
- split: validation
|
100 |
path: full/validation-*
|
101 |
default: true
|
|
|
|
|
|
|
|
|
|
|
|
|
102 |
---
|
103 |
|
104 |
# Dataset Card for AmbigQA: Answering Ambiguous Open-domain Questions
|
|
|
84 |
sequence: string
|
85 |
splits:
|
86 |
- name: train
|
87 |
+
num_bytes: 2739628
|
88 |
num_examples: 10036
|
89 |
- name: validation
|
90 |
+
num_bytes: 805756
|
91 |
num_examples: 2002
|
92 |
+
download_size: 1777867
|
93 |
+
dataset_size: 3545384
|
94 |
configs:
|
95 |
- config_name: full
|
96 |
data_files:
|
|
|
99 |
- split: validation
|
100 |
path: full/validation-*
|
101 |
default: true
|
102 |
+
- config_name: light
|
103 |
+
data_files:
|
104 |
+
- split: train
|
105 |
+
path: light/train-*
|
106 |
+
- split: validation
|
107 |
+
path: light/validation-*
|
108 |
---
|
109 |
|
110 |
# Dataset Card for AmbigQA: Answering Ambiguous Open-domain Questions
|
dataset_infos.json
CHANGED
@@ -7,66 +7,50 @@
|
|
7 |
"features": {
|
8 |
"id": {
|
9 |
"dtype": "string",
|
10 |
-
"id": null,
|
11 |
"_type": "Value"
|
12 |
},
|
13 |
"question": {
|
14 |
"dtype": "string",
|
15 |
-
"id": null,
|
16 |
"_type": "Value"
|
17 |
},
|
18 |
"annotations": {
|
19 |
"feature": {
|
20 |
"type": {
|
21 |
"dtype": "string",
|
22 |
-
"id": null,
|
23 |
"_type": "Value"
|
24 |
},
|
25 |
"answer": {
|
26 |
"feature": {
|
27 |
"dtype": "string",
|
28 |
-
"id": null,
|
29 |
"_type": "Value"
|
30 |
},
|
31 |
-
"length": -1,
|
32 |
-
"id": null,
|
33 |
"_type": "Sequence"
|
34 |
},
|
35 |
"qaPairs": {
|
36 |
"feature": {
|
37 |
"question": {
|
38 |
"dtype": "string",
|
39 |
-
"id": null,
|
40 |
"_type": "Value"
|
41 |
},
|
42 |
"answer": {
|
43 |
"feature": {
|
44 |
"dtype": "string",
|
45 |
-
"id": null,
|
46 |
"_type": "Value"
|
47 |
},
|
48 |
-
"length": -1,
|
49 |
-
"id": null,
|
50 |
"_type": "Sequence"
|
51 |
}
|
52 |
},
|
53 |
-
"length": -1,
|
54 |
-
"id": null,
|
55 |
"_type": "Sequence"
|
56 |
}
|
57 |
},
|
58 |
-
"length": -1,
|
59 |
-
"id": null,
|
60 |
"_type": "Sequence"
|
61 |
}
|
62 |
},
|
63 |
-
"
|
64 |
-
"
|
65 |
-
"builder_name": "ambig_qa",
|
66 |
"config_name": "light",
|
67 |
"version": {
|
68 |
"version_str": "1.0.0",
|
69 |
-
"description": null,
|
70 |
"major": 1,
|
71 |
"minor": 0,
|
72 |
"patch": 0
|
@@ -74,31 +58,20 @@
|
|
74 |
"splits": {
|
75 |
"train": {
|
76 |
"name": "train",
|
77 |
-
"num_bytes":
|
78 |
"num_examples": 10036,
|
79 |
-
"dataset_name":
|
80 |
},
|
81 |
"validation": {
|
82 |
"name": "validation",
|
83 |
-
"num_bytes":
|
84 |
"num_examples": 2002,
|
85 |
-
"dataset_name":
|
86 |
-
}
|
87 |
-
},
|
88 |
-
"download_checksums": {
|
89 |
-
"https://nlp.cs.washington.edu/ambigqa/data/ambignq_light.zip": {
|
90 |
-
"num_bytes": 1061383,
|
91 |
-
"checksum": "3f5dada69dec05cef1533a64945cd7bafde1aa94b0cdd6fa9a22f881206220db"
|
92 |
-
},
|
93 |
-
"https://nlp.cs.washington.edu/ambigqa/data/ambignq.zip": {
|
94 |
-
"num_bytes": 18639517,
|
95 |
-
"checksum": "e85cec5909f076c6f584322c7f05cae44dcacaec93758c110a26fcceaa8da0ce"
|
96 |
}
|
97 |
},
|
98 |
-
"download_size":
|
99 |
-
"
|
100 |
-
"
|
101 |
-
"size_in_bytes": 23246440
|
102 |
},
|
103 |
"full": {
|
104 |
"description": "AmbigNQ, a dataset covering 14,042 questions from NQ-open, an existing open-domain QA benchmark. We find that over half of the questions in NQ-open are ambiguous. The types of ambiguity are diverse and sometimes subtle, many of which are only apparent after examining evidence provided by a very large text corpus. AMBIGNQ, a dataset with\n14,042 annotations on NQ-OPEN questions containing diverse types of ambiguity.\nWe provide two distributions of our new dataset AmbigNQ: a full version with all annotation metadata and a light version with only inputs and outputs.\n",
|
|
|
7 |
"features": {
|
8 |
"id": {
|
9 |
"dtype": "string",
|
|
|
10 |
"_type": "Value"
|
11 |
},
|
12 |
"question": {
|
13 |
"dtype": "string",
|
|
|
14 |
"_type": "Value"
|
15 |
},
|
16 |
"annotations": {
|
17 |
"feature": {
|
18 |
"type": {
|
19 |
"dtype": "string",
|
|
|
20 |
"_type": "Value"
|
21 |
},
|
22 |
"answer": {
|
23 |
"feature": {
|
24 |
"dtype": "string",
|
|
|
25 |
"_type": "Value"
|
26 |
},
|
|
|
|
|
27 |
"_type": "Sequence"
|
28 |
},
|
29 |
"qaPairs": {
|
30 |
"feature": {
|
31 |
"question": {
|
32 |
"dtype": "string",
|
|
|
33 |
"_type": "Value"
|
34 |
},
|
35 |
"answer": {
|
36 |
"feature": {
|
37 |
"dtype": "string",
|
|
|
38 |
"_type": "Value"
|
39 |
},
|
|
|
|
|
40 |
"_type": "Sequence"
|
41 |
}
|
42 |
},
|
|
|
|
|
43 |
"_type": "Sequence"
|
44 |
}
|
45 |
},
|
|
|
|
|
46 |
"_type": "Sequence"
|
47 |
}
|
48 |
},
|
49 |
+
"builder_name": "parquet",
|
50 |
+
"dataset_name": "ambig_qa",
|
|
|
51 |
"config_name": "light",
|
52 |
"version": {
|
53 |
"version_str": "1.0.0",
|
|
|
54 |
"major": 1,
|
55 |
"minor": 0,
|
56 |
"patch": 0
|
|
|
58 |
"splits": {
|
59 |
"train": {
|
60 |
"name": "train",
|
61 |
+
"num_bytes": 2739628,
|
62 |
"num_examples": 10036,
|
63 |
+
"dataset_name": null
|
64 |
},
|
65 |
"validation": {
|
66 |
"name": "validation",
|
67 |
+
"num_bytes": 805756,
|
68 |
"num_examples": 2002,
|
69 |
+
"dataset_name": null
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
70 |
}
|
71 |
},
|
72 |
+
"download_size": 1777867,
|
73 |
+
"dataset_size": 3545384,
|
74 |
+
"size_in_bytes": 5323251
|
|
|
75 |
},
|
76 |
"full": {
|
77 |
"description": "AmbigNQ, a dataset covering 14,042 questions from NQ-open, an existing open-domain QA benchmark. We find that over half of the questions in NQ-open are ambiguous. The types of ambiguity are diverse and sometimes subtle, many of which are only apparent after examining evidence provided by a very large text corpus. AMBIGNQ, a dataset with\n14,042 annotations on NQ-OPEN questions containing diverse types of ambiguity.\nWe provide two distributions of our new dataset AmbigNQ: a full version with all annotation metadata and a light version with only inputs and outputs.\n",
|
light/train-00000-of-00001.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7876a94a06706dd904b32f66aa0883301c6b3906991cb5fb4db7d820f96a7c95
|
3 |
+
size 1411343
|
light/validation-00000-of-00001.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bdb0ad1c39e6df47372d38c9b82505d8d790053cd736d497fb7c387c2223121f
|
3 |
+
size 366524
|