README.md

metadata

task_categories:
  - image-classification

AutoTrain Dataset for project: weather-classification

Dataset Description

This dataset has been automatically processed by AutoTrain for project weather-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": "<771x514 RGB PIL image>",
    "target": 2
  },
  {
    "image": "<269x254 RGB PIL image>",
    "target": 8
  }
]

Dataset Fields

The dataset has the following fields (also called "features"):

{
  "image": "Image(decode=True, id=None)",
  "target": "ClassLabel(names=['dew', 'fogsmog', 'frost', 'glaze', 'hail', 'lightning', 'rain', 'rainbow', 'rime', 'sandstorm', 'snow'], 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	5484
valid	1378