Triangle104/cognitivecomputations-SexDrugsAndRockAndRoll

Sex, Drugs, and Rock and Roll Dataset

Overview

The Sex, Drugs, and Rock and Roll dataset is a comprehensive collection of controversial and adult-themed images designed to train multimodal models. The primary purpose of this dataset is to enable models to accurately detect, categorize, and manage adult-themed content, providing system implementers with the ability to filter data at a highly detailed level.

We are working on a Dolphin Vision model. This dataset will be used to train that model.

Description

The dataset contains a variety of images that fall into categories such as adult content, drug-related imagery, and themes associated with rock and roll culture. Each image is labeled with detailed metadata to facilitate training and filtering by multimodal models. This includes annotations for explicit content, drug references, nudity, violence, and other relevant attributes.

Categories

1. Adult Content

   • Nudity
   • Sexual acts
   • Explicit gestures

2. Drug-related Imagery

   • Drug use
   • Paraphernalia
   • Scenes involving intoxication

3. Rock and Roll Themes

   • Concert scenes
   • Musicians and bands
   • Symbolism associated with rock culture

Purpose and Use Cases

The dataset is intended for training AI models to:

1. Detect and categorize adult-themed content: Equip models to identify various forms of explicit content with high accuracy.
2. Enable detailed content filtering: Allow system implementers to set granular filtering criteria based on specific types of adult content.
3. Support research in content moderation: Aid researchers in developing more effective content moderation tools for social media, streaming platforms, and other online services.

Dataset Contents

The dataset contains labelled images in the following categories:

• adult_content
  • nudity
  • sexual_acts
  • explicit_gestures
• drug_related
  • drug_use
  • paraphernalia
  • intoxication
• rock_and_roll
  • concert_scenes
  • musicians
  • symbolism

Contributing

We welcome contributions to improve the dataset. If you have suggestions, new images, or metadata corrections, please submit a pull request or contact Your Mom.

Acknowledgments

We thank the contributors and organizations that provided images and annotations for this dataset.

Contact

For questions or further information, please contact Your Mom.