metadata
task_categories:
- token-classification
AutoTrain Dataset for project: code-mixed-language-identification
Dataset Description
This dataset has been automatically processed by AutoTrain for project code-mixed-language-identification.
Languages
The BCP-47 code for the dataset's language is unk.
Dataset Structure
Data Instances
A sample from this dataset looks as follows:
[
{
"feat_Unnamed: 0": 9180,
"tokens": [
"keren",
"banget",
"projectnya",
"\ud83d\udc4f"
],
"tags": [
1,
1,
3,
6
]
},
{
"feat_Unnamed: 0": 1911,
"tokens": [
"ygshop",
"juga",
"lolaaaa",
".",
"mau",
"order",
"piye",
"lek",
"loadinge",
"ngene",
"!",
"ate",
"tuku",
"piye",
"lek",
"loadingnge",
"ngene",
"?",
"welcome",
"back",
"albummmmm",
"!!"
],
"tags": [
3,
1,
3,
6,
1,
1,
2,
2,
5,
2,
6,
2,
2,
2,
2,
5,
2,
6,
0,
0,
0,
6
]
}
]
Dataset Fields
The dataset has the following fields (also called "features"):
{
"feat_Unnamed: 0": "Value(dtype='int64', id=None)",
"tokens": "Sequence(feature=Value(dtype='string', id=None), length=-1, id=None)",
"tags": "Sequence(feature=ClassLabel(names=['EN', 'ID', 'JV', 'MIX_ID_EN', 'MIX_ID_JV', 'MIX_JV_EN', 'OTH'], id=None), length=-1, id=None)"
}
Dataset Splits
This dataset is split into a train and validation split. The split sizes are as follow:
| Split name | Num samples |
|---|---|
| train | 7868 |
| valid | 3151 |