Datasets:
Modalities:
Text
Size:
10K - 100K
Upload README.md
Browse files
README.md
CHANGED
@@ -1,18 +1,18 @@
|
|
1 |
# BAYƐLƐMABAGA: Parallel French - Bambara Dataset for Machine Learning
|
2 |
|
3 |
## Overview
|
4 |
-
The Bayelemabaga dataset is a collection of 44562 aligned machine translation ready Bambara-French lines, originating from [Corpus Bambara de Reference](http://cormande.huma-num.fr/corbama/run.cgi/first_form). The dataset is constitued of text extracted from **
|
5 |
|
6 |
## Snapshot: 46976
|
7 |
| | |
|
8 |
|:---|---:|
|
9 |
| **Lines** | **46976** |
|
10 |
-
| French Tokens |
|
11 |
-
| Bambara Tokens |
|
12 |
-
| French Types |
|
13 |
-
| Bambara Types |
|
14 |
-
| Avg. Fr line length |
|
15 |
-
| Avg. Bam line length |
|
16 |
| Number of text sources | 264 |
|
17 |
|
18 |
## Data Splits
|
@@ -35,8 +35,12 @@ The Bayelemabaga dataset is a collection of 44562 aligned machine translation re
|
|
35 |
|
36 |
- `Coming soon`
|
37 |
|
|
|
|
|
|
|
|
|
38 |
### To note:
|
39 |
-
- ʃ => (sh) sound: Symbol left in the dataset, although not a part of bambara orthography nor French orthography.
|
40 |
|
41 |
## License
|
42 |
|
@@ -53,14 +57,14 @@ The Bayelemabaga dataset is a collection of 44562 aligned machine translation re
|
|
53 |
title={Machine Learning Dataset Development for Manding Languages},
|
54 |
author={
|
55 |
Valentin Vydrin and
|
56 |
-
Christopher Homan and
|
57 |
-
Michael Leventhal and
|
58 |
-
Allashera Auguste Tapo and
|
59 |
-
Marco Zampieri and
|
60 |
Jean-Jacques Meric and
|
61 |
Kirill Maslinsky and
|
62 |
Andrij Rovenchak and
|
63 |
-
|
|
|
|
|
|
|
|
|
64 |
},
|
65 |
howpublished = {url{https://github.com/robotsmali-ai/datasets}},
|
66 |
year={2022}
|
@@ -69,3 +73,4 @@ The Bayelemabaga dataset is a collection of 44562 aligned machine translation re
|
|
69 |
|
70 |
## Contacts
|
71 |
- `sdiarra <at> robotsmali.org`
|
|
|
|
1 |
# BAYƐLƐMABAGA: Parallel French - Bambara Dataset for Machine Learning
|
2 |
|
3 |
## Overview
|
4 |
+
The Bayelemabaga dataset is a collection of 44562 aligned machine translation ready Bambara-French lines, originating from [Corpus Bambara de Reference](http://cormande.huma-num.fr/corbama/run.cgi/first_form). The dataset is constitued of text extracted from **264** text files, varing from periodicals, books, short stories, blog posts, part of the Bible and the Quran.
|
5 |
|
6 |
## Snapshot: 46976
|
7 |
| | |
|
8 |
|:---|---:|
|
9 |
| **Lines** | **46976** |
|
10 |
+
| French Tokens (spacy) | 691312 |
|
11 |
+
| Bambara Tokens (daba) | 660732 |
|
12 |
+
| French Types | 32018 |
|
13 |
+
| Bambara Types | 29382 |
|
14 |
+
| Avg. Fr line length | 77.6 |
|
15 |
+
| Avg. Bam line length | 61.69 |
|
16 |
| Number of text sources | 264 |
|
17 |
|
18 |
## Data Splits
|
|
|
35 |
|
36 |
- `Coming soon`
|
37 |
|
38 |
+
### Sources
|
39 |
+
|
40 |
+
- [`sources`](./bayelemabaga/sources.txt)
|
41 |
+
|
42 |
### To note:
|
43 |
+
- ʃ => (sh/shy) sound: Symbol left in the dataset, although not a part of bambara orthography nor French orthography.
|
44 |
|
45 |
## License
|
46 |
|
|
|
57 |
title={Machine Learning Dataset Development for Manding Languages},
|
58 |
author={
|
59 |
Valentin Vydrin and
|
|
|
|
|
|
|
|
|
60 |
Jean-Jacques Meric and
|
61 |
Kirill Maslinsky and
|
62 |
Andrij Rovenchak and
|
63 |
+
Allashera Auguste Tapo and
|
64 |
+
Sebastien Diarra and
|
65 |
+
Christopher Homan and
|
66 |
+
Marco Zampieri and
|
67 |
+
Michael Leventhal
|
68 |
},
|
69 |
howpublished = {url{https://github.com/robotsmali-ai/datasets}},
|
70 |
year={2022}
|
|
|
73 |
|
74 |
## Contacts
|
75 |
- `sdiarra <at> robotsmali.org`
|
76 |
+
- `aat3261 <at> rit.edu`
|