dehatebert-mono-portugese
| Property | Value |
|---|---|
| License | Apache 2.0 |
| Paper | Deep Learning Models for Multilingual Hate Speech Detection |
| Downloads | 92,169 |
| Framework | PyTorch, JAX |
What is dehatebert-mono-portugese?
dehatebert-mono-portugese is a specialized hate speech detection model designed for Portuguese language content. Built on the multilingual BERT architecture, this model represents a monolingual approach where training is conducted using Portuguese language data. It achieved a notable validation score of 0.716119 with a learning rate of 3e-5.
Implementation Details
The model is implemented using the Transformers architecture and is compatible with both PyTorch and JAX frameworks. It's fine-tuned on multilingual BERT specifically for hate speech detection tasks in Portuguese text.
- Optimized learning rate of 3e-5 for best performance
- Based on multilingual BERT architecture
- Supports inference endpoints for deployment
Core Capabilities
- Portuguese language hate speech detection
- Binary classification of text content
- Production-ready with inference endpoint support
- Trained on specialized hate speech datasets
Frequently Asked Questions
Q: What makes this model unique?
This model is specifically optimized for Portuguese hate speech detection using a monolingual training approach, despite being based on multilingual BERT. This focused training strategy allows for better performance on Portuguese content.
Q: What are the recommended use cases?
The model is ideal for content moderation systems, social media platforms, and online communities where Portuguese language content needs to be monitored for hate speech. It can be integrated into automated content filtering systems.
