| --- |
| tags: |
| - text-classification |
| - toxicity-detection |
| model-index: |
| - name: toxicity-classifier |
| results: [] |
| --- |
| |
| # Toxicity Classifier |
|
|
| ## Overview |
|
|
| This model is a fine-tuned BERT model designed for detecting toxicity in text. It classifies input text as either "toxic" or "non-toxic" based on learned patterns from a diverse dataset of online comments and discussions. The model achieves high accuracy in identifying harmful language, making it suitable for content moderation tasks. |
|
|
| ## Model Architecture |
|
|
| The model is based on the BERT (Bidirectional Encoder Representations from Transformers) architecture, specifically `bert-base-uncased`. It consists of 12 transformer layers, each with 12 attention heads and a hidden size of 768. The final layer is a classification head that outputs probabilities for the two classes: non-toxic (0) and toxic (1). |
|
|
| ## Intended Use |
|
|
| This model is intended for use in applications requiring automated toxicity detection, such as: |
| - Social media platforms for moderating user comments. |
| - Online forums to flag potentially harmful content. |
| - Customer support systems to identify abusive language in queries. |
|
|
| It can be integrated into pipelines using the Hugging Face Transformers library. Example usage: |
|
|
| ```python |
| from transformers import pipeline |
| |
| classifier = pipeline("text-classification", model="your-username/toxicity-classifier") |
| result = classifier("This is a harmful comment.") |
| print(result) |
