Open source of fine-tuned distilbert-base-uncased model - Free deployment for classifying hate and offensive content

Distilbert Base Uncased Hate Speech Offensive Train 8 3

Developed by SetFit

This model is a fine-tuned version of distilbert-base-uncased, designed for classifying hate speech and offensive content.

Downloads 17

Release Time : 3/2/2022

Model Overview

This model is a fine-tuned version of distilbert-base-uncased, primarily used to identify and classify hate speech and offensive content.

Lightweight model

Based on the DistilBERT architecture, it is more lightweight than the original BERT model, making it suitable for resource-constrained environments.

Hate speech detection

Specifically optimized for hate speech and offensive content, effectively identifying such texts.

Text classification

Hate speech detection

Offensive content identification

Content moderation

Social media content filtering

Used to automatically detect and filter hate speech and offensive content on social media.

Accuracy: 0.549

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.1073	1.0	5	1.1393	0.0
1.0392	2.0	10	1.1729	0.0
1.0302	3.0	15	1.1694	0.2
0.9176	4.0	20	1.1846	0.2
0.8339	5.0	25	1.1663	0.2
0.7533	6.0	30	1.1513	0.4
0.6327	7.0	35	1.1474	0.4
0.4402	8.0	40	1.1385	0.4
0.3752	9.0	45	1.0965	0.2
0.3448	10.0	50	1.0357	0.2
0.2582	11.0	55	1.0438	0.2
0.1903	12.0	60	1.0561	0.2
0.1479	13.0	65	1.0569	0.2
0.1129	14.0	70	1.0455	0.2
0.1071	15.0	75	1.0416	0.4
0.0672	16.0	80	1.1164	0.4
0.0561	17.0	85	1.1846	0.6
0.0463	18.0	90	1.2040	0.6
0.0431	19.0	95	1.2078	0.6
0.0314	20.0	100	1.2368	0.6

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base