metadata
task_categories:
- image-classification
AutoTrain Dataset for project: img-classification
Dataset Description
This dataset has been automatically processed by AutoTrain for project img-classification.
Languages
The BCP-47 code for the dataset's language is unk.
Dataset Structure
Data Instances
A sample from this dataset looks as follows:
[
{
"image": "<222x163 RGB PIL image>",
"target": 0
},
{
"image": "<222x163 RGB PIL image>",
"target": 3
}
]
Dataset Fields
The dataset has the following fields (also called "features"):
{
"image": "Image(decode=True, id=None)",
"target": "ClassLabel(names=['elf', 'goblin', 'knight', 'zombie'], 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 | 1160 |
| valid | 291 |