metadata
task_categories:
- image-classification
AutoTrain Dataset for project: model_me
Dataset Description
This dataset has been automatically processed by AutoTrain for project model_me.
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": "<600x1100 RGB PIL image>",
"target": 7
},
{
"image": "<1000x750 RGBA PIL image>",
"target": 0
}
]
Dataset Fields
The dataset has the following fields (also called "features"):
{
"image": "Image(decode=True, id=None)",
"target": "ClassLabel(num_classes=8, names=['3d', 'above', 'front', 'game', 'ilustration', 'inside', 'object', 'pixel art'], 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 | 468 |
| valid | 121 |