DAIE

URL: https://github.com/Jorffy/DAIE

Description: Official implementation of the paper “Dual-Level Adaptive Incongruity-Enhanced Model for Multimodal Sarcasm Detection”, accepted in Neurocomputing.

Project Overview

DAIE is a multimodal sarcasm detection model designed to process and analyze sarcasm using multiple data modalities, such as text and images.

Training Methods:

• The model is trained on a publicly available multimodal sarcasm detection dataset.
• Uses adaptive incongruity enhancement for improved sarcasm detection.
• Requires dataset preprocessing to align with previous work.
• Custom dataset paths must be set in main.py (DATA_PATH and IMG_PATH).

Results:

• No explicit model performance metrics or evaluation results are provided in the repository.

Dataset:

• Dataset is not directly included but can be obtained from:
  https://github.com/headacheboy/data-of-multimodal-sarcasm-detection.