README.md

metadata

title: Toxic Speech Classifier
emoji: 🤖
colorFrom: red
colorTo: yellow
sdk: gradio
sdk_version: 6.6.0
app_file: app.py
pinned: false

Toxic Speech Classifier 🤖

A fine-tuned Gemma3-1B model for detecting and classifying toxic, insulting, or harmful language in text.

Description

This demo uses a fine-tuned language model to analyze text and determine whether it contains toxic or harmful content. The model returns structured information including toxicity label, tags, severity, and reasoning.

Usage

Simply enter any text, and the model will classify whether it is toxic or non-toxic, along with detailed structured output.

Examples

• "You are absolutely worthless and no one will ever love you."
• "Shut up you brainless moron, nobody asked for your stupid opinion."
• "The weather today is really nice, I enjoyed my walk in the park."
• "Thank you for your help, I really appreciate everything you did."

Model

• Base Model: Gemma3-1B
• Fine-tuned on: Insult and toxic speech classification data
• Model ID: berkeruveyik/toxic-speech-finetune-with-gemma-3-1b-v1

Output Fields

Field	Description
is_toxic	Whether the text is toxic (true/false)
label	Classification label (e.g., insult, threat, neutral)
tags	Relevant tags describing the type of toxicity
reason	Explanation for the classification
severity	Severity level of the toxic content

License

Please check the model license on the Hugging Face model page.