severo HF staff commited on
Commit
6761d73
0 Parent(s):

Duplicate from severo/danish-wit

Browse files
.gitattributes ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ *.7z filter=lfs diff=lfs merge=lfs -text
2
+ *.arrow filter=lfs diff=lfs merge=lfs -text
3
+ *.bin filter=lfs diff=lfs merge=lfs -text
4
+ *.bz2 filter=lfs diff=lfs merge=lfs -text
5
+ *.ckpt filter=lfs diff=lfs merge=lfs -text
6
+ *.ftz filter=lfs diff=lfs merge=lfs -text
7
+ *.gz filter=lfs diff=lfs merge=lfs -text
8
+ *.h5 filter=lfs diff=lfs merge=lfs -text
9
+ *.joblib filter=lfs diff=lfs merge=lfs -text
10
+ *.lfs.* filter=lfs diff=lfs merge=lfs -text
11
+ *.lz4 filter=lfs diff=lfs merge=lfs -text
12
+ *.mlmodel filter=lfs diff=lfs merge=lfs -text
13
+ *.model filter=lfs diff=lfs merge=lfs -text
14
+ *.msgpack filter=lfs diff=lfs merge=lfs -text
15
+ *.npy filter=lfs diff=lfs merge=lfs -text
16
+ *.npz filter=lfs diff=lfs merge=lfs -text
17
+ *.onnx filter=lfs diff=lfs merge=lfs -text
18
+ *.ot filter=lfs diff=lfs merge=lfs -text
19
+ *.parquet filter=lfs diff=lfs merge=lfs -text
20
+ *.pb filter=lfs diff=lfs merge=lfs -text
21
+ *.pickle filter=lfs diff=lfs merge=lfs -text
22
+ *.pkl filter=lfs diff=lfs merge=lfs -text
23
+ *.pt filter=lfs diff=lfs merge=lfs -text
24
+ *.pth filter=lfs diff=lfs merge=lfs -text
25
+ *.rar filter=lfs diff=lfs merge=lfs -text
26
+ *.safetensors filter=lfs diff=lfs merge=lfs -text
27
+ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
28
+ *.tar.* filter=lfs diff=lfs merge=lfs -text
29
+ *.tflite filter=lfs diff=lfs merge=lfs -text
30
+ *.tgz filter=lfs diff=lfs merge=lfs -text
31
+ *.wasm filter=lfs diff=lfs merge=lfs -text
32
+ *.xz filter=lfs diff=lfs merge=lfs -text
33
+ *.zip filter=lfs diff=lfs merge=lfs -text
34
+ *.zst filter=lfs diff=lfs merge=lfs -text
35
+ *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ # Audio files - uncompressed
37
+ *.pcm filter=lfs diff=lfs merge=lfs -text
38
+ *.sam filter=lfs diff=lfs merge=lfs -text
39
+ *.raw filter=lfs diff=lfs merge=lfs -text
40
+ # Audio files - compressed
41
+ *.aac filter=lfs diff=lfs merge=lfs -text
42
+ *.flac filter=lfs diff=lfs merge=lfs -text
43
+ *.mp3 filter=lfs diff=lfs merge=lfs -text
44
+ *.ogg filter=lfs diff=lfs merge=lfs -text
45
+ *.wav filter=lfs diff=lfs merge=lfs -text
46
+ # Image files - uncompressed
47
+ *.bmp filter=lfs diff=lfs merge=lfs -text
48
+ *.gif filter=lfs diff=lfs merge=lfs -text
49
+ *.png filter=lfs diff=lfs merge=lfs -text
50
+ *.tiff filter=lfs diff=lfs merge=lfs -text
51
+ # Image files - compressed
52
+ *.jpg filter=lfs diff=lfs merge=lfs -text
53
+ *.jpeg filter=lfs diff=lfs merge=lfs -text
54
+ *.webp filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,158 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ pretty_name: Danish WIT
3
+ language:
4
+ - da
5
+ license:
6
+ - cc-by-sa-4.0
7
+ size_categories:
8
+ - 100K<n<1M
9
+ source_datasets:
10
+ - wikimedia/wit_base
11
+ task_categories:
12
+ - image-to-text
13
+ - zero-shot-image-classification
14
+ - feature-extraction
15
+ task_ids:
16
+ - image-captioning
17
+ ---
18
+
19
+ # Dataset Card for Danish WIT
20
+
21
+ ## Dataset Description
22
+
23
+ - **Repository:** <https://gist.github.com/saattrupdan/bb6c9c52d9f4b35258db2b2456d31224>
24
+ - **Point of Contact:** [Dan Saattrup Nielsen](mailto:[email protected])
25
+ - **Size of downloaded dataset files:** 7.5 GB
26
+ - **Size of the generated dataset:** 7.8 GB
27
+ - **Total amount of disk used:** 15.3 GB
28
+
29
+ ### Dataset Summary
30
+
31
+ Google presented the Wikipedia Image Text (WIT) dataset in [July
32
+ 2021](https://dl.acm.org/doi/abs/10.1145/3404835.3463257), a dataset which contains
33
+ scraped images from Wikipedia along with their descriptions. WikiMedia released
34
+ WIT-Base in [September
35
+ 2021](https://techblog.wikimedia.org/2021/09/09/the-wikipedia-image-caption-matching-challenge-and-a-huge-release-of-image-data-for-research/),
36
+ being a modified version of WIT where they have removed the images with empty
37
+ "reference descriptions", as well as removing images where a person's face covers more
38
+ than 10% of the image surface, along with inappropriate images that are candidate for
39
+ deletion. This dataset is the Danish portion of the WIT-Base dataset, consisting of
40
+ roughly 160,000 images with associated Danish descriptions. We release the dataset
41
+ under the [CC BY-SA 4.0 license](https://creativecommons.org/licenses/by-sa/4.0/), in
42
+ accordance with WIT-Base's [identical
43
+ license](https://huggingface.co/datasets/wikimedia/wit_base#licensing-information).
44
+
45
+
46
+ ### Supported Tasks and Leaderboards
47
+
48
+ Training machine learning models for caption generation, zero-shot image classification
49
+ and text-image search are the intended tasks for this dataset. No leaderboard is active
50
+ at this point.
51
+
52
+
53
+ ### Languages
54
+
55
+ The dataset is available in Danish (`da`).
56
+
57
+
58
+ ## Dataset Structure
59
+
60
+ ### Data Instances
61
+
62
+ - **Size of downloaded dataset files:** 7.5 GB
63
+ - **Size of the generated dataset:** 7.8 GB
64
+ - **Total amount of disk used:** 15.3 GB
65
+
66
+ An example from the `train` split looks as follows.
67
+ ```
68
+ {
69
+ "image": {
70
+ "bytes": b"\xff\xd8\xff\xe0\x00\x10JFIF...",
71
+ "path": None
72
+ },
73
+ "image_url": "https://upload.wikimedia.org/wikipedia/commons/4/45/Bispen_-_inside.jpg",
74
+ "embedding": [2.8568285, 2.9562542, 0.33794892, 8.753725, ...],
75
+ "metadata_url": "http://commons.wikimedia.org/wiki/File:Bispen_-_inside.jpg",
76
+ "original_height": 3161,
77
+ "original_width": 2316,
78
+ "mime_type": "image/jpeg",
79
+ "caption_attribution_description": "Kulturhuset Bispen set indefra. Biblioteket er til venstre",
80
+ "page_url": "https://da.wikipedia.org/wiki/Bispen",
81
+ "attribution_passes_lang_id": True,
82
+ "caption_alt_text_description": None,
83
+ "caption_reference_description": "Bispen set indefra fra 1. sal, hvor ....",
84
+ "caption_title_and_reference_description": "Bispen [SEP] Bispen set indefra ...",
85
+ "context_page_description": "Bispen er navnet på det offentlige kulturhus i ...",
86
+ "context_section_description": "Bispen er navnet på det offentlige kulturhus i ...",
87
+ "hierarchical_section_title": "Bispen",
88
+ "is_main_image": True,
89
+ "page_changed_recently": True,
90
+ "page_title": "Bispen",
91
+ "section_title": None
92
+ }
93
+ ```
94
+
95
+ ### Data Fields
96
+
97
+ The data fields are the same among all splits.
98
+
99
+ - `image`: a `dict` feature.
100
+ - `image_url`: a `str` feature.
101
+ - `embedding`: a `list` feature.
102
+ - `metadata_url`: a `str` feature.
103
+ - `original_height`: an `int` or `NaN` feature.
104
+ - `original_width`: an `int` or `NaN` feature.
105
+ - `mime_type`: a `str` or `None` feature.
106
+ - `caption_attribution_description`: a `str` or `None` feature.
107
+ - `page_url`: a `str` feature.
108
+ - `attribution_passes_lang_id`: a `bool` or `None` feature.
109
+ - `caption_alt_text_description`: a `str` or `None` feature.
110
+ - `caption_reference_description`: a `str` or `None` feature.
111
+ - `caption_title_and_reference_description`: a `str` or `None` feature.
112
+ - `context_page_description`: a `str` or `None` feature.
113
+ - `context_section_description`: a `str` or `None` feature.
114
+ - `hierarchical_section_title`: a `str` feature.
115
+ - `is_main_image`: a `bool` or `None` feature.
116
+ - `page_changed_recently`: a `bool` or `None` feature.
117
+ - `page_title`: a `str` feature.
118
+ - `section_title`: a `str` or `None` feature.
119
+
120
+ ### Data Splits
121
+
122
+ Roughly 2.60% of the WIT-Base dataset comes from the Danish Wikipedia. We have split
123
+ the resulting 168,740 samples into a training set, validation set and testing set of
124
+ the following sizes:
125
+
126
+ | split | samples |
127
+ |---------|--------:|
128
+ | train | 167,460 |
129
+ | val | 256 |
130
+ | test | 1,024 |
131
+
132
+
133
+ ## Dataset Creation
134
+
135
+ ### Curation Rationale
136
+
137
+ It is quite cumbersome to extract the Danish portion of the WIT-Base dataset,
138
+ especially as the dataset takes up 333 GB of disk space, so the curation of Danish-WIT
139
+ is purely to make it easier to work with the Danish portion of it.
140
+
141
+ ### Source Data
142
+
143
+ The original data was collected from WikiMedia's
144
+ [WIT-Base](https://huggingface.co/datasets/wikimedia/wit_base) dataset, which in turn
145
+ comes from Google's [WIT](https://huggingface.co/datasets/google/wit) dataset.
146
+
147
+
148
+ ## Additional Information
149
+
150
+ ### Dataset Curators
151
+
152
+ [Dan Saattrup Nielsen](https://saattrupdan.github.io/) from the [The Alexandra
153
+ Institute](https://alexandra.dk/) curated this dataset.
154
+
155
+ ### Licensing Information
156
+
157
+ The dataset is licensed under the [CC BY-SA 4.0
158
+ license](https://creativecommons.org/licenses/by-sa/4.0/).
data/test-00000-of-00001-3ec8196e06df8f4c.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8c02c2b7821950bd9b6ce4bdf266bf4e9e40245fb479310f07ef40bcf1527b10
3
+ size 48690082
data/train-00000-of-00017-aa475bcd75eefc19.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3f4d5ab045c5e692d7dd0fad32343e7577fdb0ee1f00bdf84230c79a270bf41d
3
+ size 465600155
data/train-00001-of-00017-1225887439989ed4.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:26259a92b6464d76ebd65b9b687d2afddb178adbb93582b1766cbd1b68af4ce4
3
+ size 465630667
data/train-00002-of-00017-667c45a8c6767457.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:851207234e2a6eb6f761de24e6566a95f824ad2e954b3e9e3ac03140158f9119
3
+ size 463791591
data/train-00003-of-00017-82347324c58594a8.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a27f5168d2a7785928261ff1dc7315ad05d72186533710a215c01f8888e4ed55
3
+ size 456174885
data/train-00004-of-00017-8454ae2e6039bf5d.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2217e783ad73900743dd72e8de0b6865dd0aa8cbca721fd33cb0b0dd38614e27
3
+ size 465359002
data/train-00005-of-00017-cf8c79d5dbfda653.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e4e413df8e2ccd5a81e3274e9b7dea3ef404e86ed329d3a6c9749fab99d2a56b
3
+ size 469056951
data/train-00006-of-00017-a83fb7aaa888b95c.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3938000df6650c0c00aad20dbde056b779130e55804af193a48137839d6d49ac
3
+ size 460266715
data/train-00007-of-00017-67e71c760dfae1f4.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b315e8998b1abfbeb58779381c63d0b7dc04d081ac8d7385ec716de1b7676a2
3
+ size 468290766
data/train-00008-of-00017-c447248912541b57.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6bb76e5382db576bfae0ab6c84a5680ad6acc7062163c862040ed85bc0a7f558
3
+ size 490068796
data/train-00009-of-00017-1d86fe3576792ab9.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:adaec47341768d79e0be5afd808cfc7a9993ca73311b4d961efb9619214d9dd3
3
+ size 460388234
data/train-00010-of-00017-8b1b99bc37342ce9.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b6e1e428e5b1608683e67479fba172e6f6147efb5be7007e902417df946b25d
3
+ size 476461776
data/train-00011-of-00017-9f1f1d9aa3f781d1.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2a2d0f63a78067f895c03ef91e239f5b160b5f81e3ad9aaaff16456a5d675d33
3
+ size 470328807
data/train-00012-of-00017-97ea039e02b99866.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f2119c9cc07ce2e129bcb74f67f26526fecc42a58234464c03bef9ac9bb5a3d9
3
+ size 457166707
data/train-00013-of-00017-7c352901296fb935.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2de6c6f976eaa776460d21f9fd0c8aa120eea1c8bb65d255c07825ad9fa66244
3
+ size 464411231
data/train-00014-of-00017-7e34b1e285827eac.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:af21d1e0118c8d12d071c2ad45d6fa7fd7cac5efd81dcbdfd82c92a2641435fa
3
+ size 466562374
data/train-00015-of-00017-fd451bfc4a44f7c9.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:238dd5f69788c7a52953ccb311fc5df7e91b143c6c12593a28c834e5efc0a6da
3
+ size 460365864
data/train-00016-of-00017-949c84a01c9b6ab5.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:880a85295dee9226f0cdee39288af78baac87f113aae6a8bb24b22f9e4888f92
3
+ size 480453069
data/val-00000-of-00001-4aebc80d34057809.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:266304cd7999dfd7b74373a52db8061acee125c606fc9caa5a5a3b7e770bdaab
3
+ size 11429349