docs/features/text_classifier.md

English Text Classifier

Overview

Detects whether English-language text is AI-generated or human-written.

Model Pipeline

• Tokenizer: GPT-2 Tokenizer
• Model: Custom trained binary classifier

Dataset

• Balanced dataset: Human vs AI-generated (ChatGPT, Claude, etc.)
• Tokenized and fed into the model using PyTorch/TensorFlow

Code Location

• Controller: features/text_classifier/controller.py
• Inference: features/text_classifier/inferencer.py
• Model Loader: features/text_classifier/model_loader.py
• Preprocessor: features/text_classifier/preprocess.py

API

• Endpoint: ENDPOINTS
• Input: Raw English text
• Output: Prediction result with probability/confidence

🔙 Back to Main README