FaiyazAzam/hw1-image-ds-groot-224

Dataset Card: [BabyGroot-Image]

Purpose

This dataset was created for educational purposes as part of a homework assignment.
It is intended to practice building, augmenting, and publishing datasets on the Hugging Face Hub.

Composition

• Number of samples:
  • Original: N samples
  • Augmented: M samples
• Features:
  • [list features/columns here, e.g. CarModel, Year, Mileage... or text, complexity_label... or image, has_groot]

Collection

• Source:
  • Images (Baby Groot): photos taken with a personal camera/phone (no people, no PII).
• Size:
  • At least 30 original entries

Preprocessing & Augmentation

• Preprocessing:
  • Images: resized to 224x224, converted to RGB.
• Augmentation techniques:
  • Images: flips, rotations, affine transforms, perspective, color jitter, Gaussian blur, noise, cutout.

Labels

• Images: Target = has_groot (binary, yes/no).

Splits

• original: manually collected data (30+).
• augmented: synthetic data (≥300 for images).

Intended Use & Limitations

• Use: Coursework, experimenting with dataset creation, augmentation, and Hugging Face tools.
• Limits:
  • Not suitable for production ML training.
  • Synthetic samples may not represent real-world distributions.
  • Small dataset sizes → limited generalization.

Ethical Notes

• No personally identifiable information (PII) or sensitive content.
• All images are of inanimate objects (a Baby Groot figure + random desk items).
• Labels are for demonstration purposes only.

License

MIT License (for educational use).

AI Usage Disclosure

• No generative AI was used to produce original image samples; only augmentation steps.