Spaces:
Running
Running
File size: 1,593 Bytes
8cc11fa e44075f 8cc11fa e44075f 8cc11fa e44075f 8cc11fa e44075f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 |
---
title: Emoji Offensive Agent
emoji: 🧠
colorFrom: pink
colorTo: yellow
sdk: streamlit
app_file: app.py
pinned: false
license: apache-2.0
short_description: Detects offensive emoji-based social media content.
---
# 🧠 Emoji Offensive Agent
This AI agent aims to detect hate speech or offensive content in social media texts, especially those that include **emojis**, **homophones**, or **subculture wordplay**.
It combines two pipelines:
1. **Emoji Translator** – Based on the fine-tuned `Qwen1.5-7B-Chat` model to convert emoji expressions into Chinese text.
2. **Hate Speech Classifier** – Uses a large pre-trained classifier (e.g., `roberta-offensive-language-detection`) to determine whether the final text is offensive (label `1`) or not (label `0`).
---
## 📦 Project Files
- `app.py`: The Streamlit app entry point.
- `agent.py`: The core logic for emoji translation + classification pipeline.
- `README.md`: You're reading it!
- `.gitattributes`: LFS file config (for large model files, if any).
---
## 🚀 Deployment
This Space uses **Streamlit** as its frontend framework. The backend pipeline automatically performs:
1. Emoji-to-text translation
2. Classification using a powerful hate speech model
You can deploy it on Hugging Face Spaces by making sure:
- The `app.py` file exists in the root directory.
- This README is properly parsed (with valid YAML at the top).
---
## 📍Example
Input:
`你是 🐷`
Intermediate translation:
`你是猪鼻子`
Final classification:
**Offensive** 🔴 (`label: 1`)
---
Made with ❤️ by JenniferHJF
|