Commit
•
6249cd5
1
Parent(s):
7ee61d6
Update README.md
Browse files
README.md
CHANGED
@@ -5,12 +5,14 @@ language:
|
|
5 |
- fr
|
6 |
size_categories:
|
7 |
- 100K<n<1M
|
|
|
8 |
---
|
9 |
|
10 |
-
#
|
11 |
-
|
12 |
-
|
13 |
-
|
|
|
14 |
|
15 |
|
16 |
## Usage
|
@@ -23,11 +25,7 @@ dataset
|
|
23 |
DatasetDict({
|
24 |
train: Dataset({
|
25 |
features: ['context', 'question', 'answer', 'answer_start', 'dataset'],
|
26 |
-
num_rows:
|
27 |
-
})
|
28 |
-
test: Dataset({
|
29 |
-
features: ['context', 'question', 'answer', 'answer_start', 'dataset'],
|
30 |
-
num_rows: 3188
|
31 |
})
|
32 |
validation: Dataset({
|
33 |
features: ['context', 'question', 'answer', 'answer_start', 'dataset'],
|
@@ -48,27 +46,61 @@ dataset_train.head()
|
|
48 |
4 Beyoncé Giselle Knowles-Carter (/ biːˈjɒnseɪ /... Combien de Grammy Awards Beyoncé a-t-elle gagn... cinq 629 pragnakalp/squad_v2_french_translated
|
49 |
```
|
50 |
|
51 |
-
##
|
52 |
-
-
|
53 |
-
-
|
54 |
-
-
|
55 |
-
-
|
56 |
-
-
|
57 |
|
58 |
## Split
|
59 |
-
-
|
60 |
-
- `validation`
|
61 |
-
- `test` correspond à l'échantillon de validation de FQuAD (le jeu de données de test de FQuAD n'étant pas dispo en open-source, le jeu de validation est devenu celui sur lequel les différents modèles se comparent) + ce même jeu de donnée augmenté au format SQuAD v2 (= fquad_v2) (aucun shuffle n'a été effectué)
|
62 |
|
63 |
## Datasets
|
64 |
|
65 |
-
| Dataset | Format | Train split | Dev split | Test split |
|
66 |
-
| ----------- | ----------- | ----------- | ----------- | ----------- |
|
67 |
-
| [piaf](https://www.data.gouv.fr/en/datasets/piaf-le-dataset-francophone-de-questions-reponses/)| SQuAD
|
68 |
-
| piaf_v2| SQuAD
|
69 |
-
| [fquad](https://fquad.illuin.tech/)| SQuAD
|
70 |
-
| fquad_v2 | SQuAD
|
71 |
-
| [lincoln/newsquadfr](https://huggingface.co/datasets/lincoln/newsquadfr) | SQuAD
|
72 |
-
| lincoln/newsquadfr_v2 | SQuAD
|
73 |
-
| [pragnakalp/squad_v2_french_translated](https://huggingface.co/datasets/pragnakalp/squad_v2_french_translated)| SQuAD
|
74 |
-
| pragnakalp/squad_v2_french_translated_v2| SQuAD
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
- fr
|
6 |
size_categories:
|
7 |
- 100K<n<1M
|
8 |
+
license: cc-by-4.0
|
9 |
---
|
10 |
|
11 |
+
# Dataset information
|
12 |
+
Dataset concatenating all QA datasets with context available in French and open-source.
|
13 |
+
In addition, an augmented version of these datasets has been added (same context but different questions to create data in SQuADv2 format).
|
14 |
+
In total, there are 221,348 training data, **910** validation data and 6,376 test data (the first 3,188 rows correspond to SQuADv2 format, the remaining 3,188 to SQuADv2 format).
|
15 |
+
In practice, due to the restrictive license for the FQUAD 1.0 dataset, we can only share **179,886** rows of the 221,348 training data and not the test dataset.
|
16 |
|
17 |
|
18 |
## Usage
|
|
|
25 |
DatasetDict({
|
26 |
train: Dataset({
|
27 |
features: ['context', 'question', 'answer', 'answer_start', 'dataset'],
|
28 |
+
num_rows: 179886
|
|
|
|
|
|
|
|
|
29 |
})
|
30 |
validation: Dataset({
|
31 |
features: ['context', 'question', 'answer', 'answer_start', 'dataset'],
|
|
|
46 |
4 Beyoncé Giselle Knowles-Carter (/ biːˈjɒnseɪ /... Combien de Grammy Awards Beyoncé a-t-elle gagn... cinq 629 pragnakalp/squad_v2_french_translated
|
47 |
```
|
48 |
|
49 |
+
## Columns
|
50 |
+
- the `context` column contains the context
|
51 |
+
- the `question` column contains the question
|
52 |
+
- the `answer` column contains the answer (has been replaced by `no_answer` for rows in SQuAD v2 format)
|
53 |
+
- the `answer_start` column contains the start position of the answer in the context (has been replaced by `-1` for rows in SQuAD v2 format)
|
54 |
+
- the `dataset` column identifies the row's original dataset (if you wish to apply filters to it, rows in SQuAD v2 format are indicated with the suffix `_v2` in the dataset name)
|
55 |
|
56 |
## Split
|
57 |
+
- train` corresponds to the concatenation of the training dataset from `pragnakalp/squad_v2_english_translated` + `lincoln/newsquadfr` + `PIAFv1.2` + the augmented version of each dataset in SQuADv2 format (no shuffle has been performed)
|
58 |
+
- `validation` corresponds to the concatenation of the newsquadfr validation dataset + this same dataset expanded in SQuAD v2 format (= newsquadfr_v2) (no shuffle performed)
|
|
|
59 |
|
60 |
## Datasets
|
61 |
|
62 |
+
| Dataset | Format | Train split | Dev split | Test split | Available in frenchQA |
|
63 |
+
| ----------- | ----------- | ----------- | ----------- | ----------- | ------------------------ |
|
64 |
+
| [piaf](https://www.data.gouv.fr/en/datasets/piaf-le-dataset-francophone-de-questions-reponses/)| SQuAD 1.0 | 9 224 Q & A | X | X | Yes |
|
65 |
+
| piaf_v2| SQuAD 2.0 | 9 224 Q & A | X | X | Yes |
|
66 |
+
| [fquad](https://fquad.illuin.tech/)| SQuAD 1.0 | 20 731 Q & A | 3 188 Q & A (is not used for training, but as a test dataset) | 2 189 Q & A (not freely available)| No due to the license |
|
67 |
+
| fquad_v2 | SQuAD 2.0 | 20 731 Q & A | 3 188 Q & A (is not used for training, but as a test dataset) | X | No due to the license |
|
68 |
+
| [lincoln/newsquadfr](https://huggingface.co/datasets/lincoln/newsquadfr) | SQuAD 1.0 | 1 650 Q & A | 455 Q & A | X | Yes |
|
69 |
+
| lincoln/newsquadfr_v2 | SQuAD 2.0 | 1 650 Q & A | 455 Q & A | X | Yes |
|
70 |
+
| [pragnakalp/squad_v2_french_translated](https://huggingface.co/datasets/pragnakalp/squad_v2_french_translated)| SQuAD 2.0 | 79 069 Q & A | X | X | Yes |
|
71 |
+
| pragnakalp/squad_v2_french_translated_v2| SQuAD 2.0 | 79 069 Q & A | X | X | Yes |
|
72 |
+
|
73 |
+
## Statistics
|
74 |
+
|
75 |
+
The question type distribution is as follows:
|
76 |
+
|
77 |
+
| Type of question | Frequency in percent |
|
78 |
+
| ----------- | ----------- |
|
79 |
+
|What (que) |55.02|
|
80 |
+
|Who (qui) |15.96|
|
81 |
+
|How much (combien)|7.92|
|
82 |
+
|When (quand) |6.90|
|
83 |
+
|Where (où) |3.15|
|
84 |
+
|How (comment) |3.76|
|
85 |
+
|What (quoi) |2.60|
|
86 |
+
|Why (pourquoi) |1.25|
|
87 |
+
|Other |3.44|
|
88 |
+
|
89 |
+
The number of questions containing a negation, e.g. "What was the name of Chopin's first music teacher who was not an amateur musician?", is estimated at 3.55% of the total questions.
|
90 |
+
|
91 |
+
|
92 |
+
For information, the distribution of the complete dataset (containing FQUAD 1.0 and FQUAD 1.0 data in SQUAD 2.0 format) is as follows:
|
93 |
+
|
94 |
+
| Type of question | Frequency in percent |
|
95 |
+
| ----------- | ----------- |
|
96 |
+
|What (que) |55.12|
|
97 |
+
|Who (qui) |16.24|
|
98 |
+
|How much (combien)|7.56|
|
99 |
+
|When (quand) |6.85|
|
100 |
+
|Where (où) |3.98|
|
101 |
+
|How (comment) |3.76|
|
102 |
+
|What (quoi) |2.94|
|
103 |
+
|Why (pourquoi) |1.41|
|
104 |
+
|Other |2.14|
|
105 |
+
|
106 |
+
The number of questions containing a negation, e.g. "What was the name of Chopin's first music teacher who was not an amateur musician?", is estimated at 3.07% of the total questions.
|