BETO Contextualized Hate Speech Detection
| Property | Value |
|---|---|
| Parameter Count | 110M |
| Model Type | Text Classification |
| Architecture | BERT (Spanish BETO) |
| Downloads | 38,841 |
| Language | Spanish |
What is beto-contextualized-hate-speech?
BETO-contextualized-hate-speech is a sophisticated Spanish language model designed for detecting hate speech in news article comments. Built on the BETO architecture (Spanish BERT), this model stands out for its ability to analyze both the comment content and its context, providing a multi-label classification across eight distinct categories of hate speech.
Implementation Details
The model implements a multi-label classification approach, processing input in the format "TEXT [SEP] CONTEXT". It utilizes the Transformers architecture with PyTorch backend and is optimized for Spanish language processing. The model evaluates content across multiple dimensions including gender, sexuality, racism, class discrimination, and more.
- Utilizes BERT-based architecture optimized for Spanish
- Supports contextual analysis through specialized input formatting
- Implements multi-label classification for granular hate speech detection
- Uses Safetensors for efficient model storage and loading
Core Capabilities
- Detection of hate speech against women, LGBTI, racial groups, and other categories
- Analysis of classist and political discrimination
- Identification of appearance-based harassment
- Detection of calls to violent action
- Context-aware classification for improved accuracy
Frequently Asked Questions
Q: What makes this model unique?
This model's unique strength lies in its ability to consider both the comment and its context for classification, providing a more nuanced understanding of hate speech. It also offers granular classification across eight distinct categories plus violence detection.
Q: What are the recommended use cases?
The model is ideal for content moderation of Spanish language news websites, social media platforms, and online forums. It's particularly useful for organizations needing to monitor and filter hate speech with consideration for context and specific target groups.
