SweeLingNiu/distilbert-intent-classifier

Model Card for Model ID

A DistilBERT-base-uncased model fine-tuned for classifying retail customer support queries into 27 intents (e.g., cancel_order, recover_password).

Achieves 99.68% test accuracy with 200ms latency, optimized for real-time deployment.

Model Details

Usage

Example Use:

from transformers import pipeline
classifier = pipeline("text-classification", model="sweelingniu/distilbert-intent-classifier")
classifier("How to reset my password?")  # Output: ('recover_password', 0.999)

DistilBERT Intent Classifier Model Card

Model Description

A DistilBERT-based model fine-tuned for intent classification tasks. Optimized for identifying user intents in conversational text with reduced computational requirements compared to BERT.

• Developed by: Sweeling Niu
• Model type: Transformer-based text classifier
• Language(s): English
• License: Apache 2.0
• Finetuned from: distilbert-base-uncased

Uses

Direct Use

• Chatbot intent detection
• Customer query classification
• Conversational AI systems

Out-of-Scope Use

• Multilingual applications
• Generative tasks
• Highly domain-specific classification without retraining

How to Use

from transformers import pipeline

classifier = pipeline(
    "text-classification",
    model="sweelingniu/distilbert-intent-classifier"
)
result = classifier("I need help resetting my password")