Skip to content

HapticGen/hapticgen-dataset

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 

Repository files navigation

HapticGen Dataset

This repository contains the dataset accompanying HapticGen: Generative Text-to-Vibration Model for Streamlining Haptic Design. For more details regarding this dataset, please refer to Appendix D, which includes the full datasheet.

Publication

Youjin Sung, Kevin John, Sang Ho Yoon, and Hasti Seifi. 2025. HapticGen: Generative Text-to-Vibration Model for Streamlining Haptic Design. In CHI Conference on Human Factors in Computing Systems (CHI ’25), April 26-May 1, 2025, Yokohama, Japan. ACM, New York, NY, USA, 23 pages. https://doi.org/10.1145/3706598.3713609

Field Descriptions

  • filename: The name of the haptic signal .wav file generated for the user.
  • user_prompt: The original natural language description of the tactile sensation provided by the user.
  • model: The model used to generate the signal. The possible models include: "HapticGen": The final, fine-tuned model. "Baseline-AudioGen": A baseline model for A/B testing. "Initial": An early version of the HapticGen model without fine-tuning.
  • vote: User feedback on the quality of the generated signal, where 1 indicates positive feedback (thumbs up), and -1 indicates negative feedback (thumbs down).
  • prompt_variant: (Optional) A variant of the original prompt provided to the model for generating variations. This was not shown to users (not considered for vote).

Schema

interface Metadata {
  /** The name of the corresponding .wav file. */
  filename: string;

  /** The natural language prompt provided by the user, describing the desired tactile sensation. */
  user_prompt: string;

  /**
   * The model used to generate the vibrotactile signal.
   * Possible values:
   * - "HapticGen": Final fine-tuned generative model.
   * - "Baseline-AudioGen": Baseline audio generation model for A/B comparison.
   * - "HapticGen-Initial": Early version of the HapticGen model (not fine-tuned, no normalization, ...)
   */
  model: "HapticGen" | "Baseline-AudioGen" | "HapticGen-Initial";

  /**
   * User feedback on the generated signal.
   * Possible values:
   * - 1: Positive feedback (thumbs up).
   * - -1: Negative feedback (thumbs down).
   */
  vote: 1 | -1;

  /**
   * (Optional) A variant of the original user prompt given to the model.
   * Used to create greater variations in the model output, but is not shown to the user.
   */
  prompt_variant?: string;
}

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Packages

No packages published