README.md

metadata

language:
  - en
license: mit
tags:
  - image-classification
  - computer-vision
task_categories:
  - image-classification
task_ids:
  - multi-class-image-classification
dataset_info:
  features:
    - name: image
      dtype:
        array3_d:
          shape:
            - 128
            - 128
            - 3
          dtype: uint8
    - name: label
      dtype:
        class_label:
          names:
            '0': cats
            '1': dogs
  splits:
    - name: train
      num_bytes: 921696000
      num_examples: 8000
    - name: test
      num_bytes: 230424000
      num_examples: 2000
  download_size: 487392383
  dataset_size: 1152120000
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*

Cats and Dogs Image Classification Dataset

This dataset contains images of cats and dogs, intended for image classification tasks. It includes two classes: "cats" and "dogs".

Dataset Structure

The dataset is structured into two splits:

• train: Contains 8000 images for training.
• test: Contains 2000 images for testing.

Images are stored in RGB format with a resolution of 128x128 pixels.

Data Loading and Usage

The dataset can be loaded using the Hugging Face Datasets library:

from datasets import load_dataset

dataset = load_dataset("cats_dogs_dataset")
This will return a DatasetDict object with the train and test splits.

Example Usage
Python

from datasets import load_dataset

dataset = load_dataset("cats_dogs_dataset")

# Access the first training example
example = dataset["train"]

# Print the image and label
print(example["image"])
print(example["label"])