Check-in or Not Classifier (v2)

🔍 Overview

The Check-in or Not Classifier is a binary text classification model designed to determine whether a given message is a valid daily check-in.

This model is part of a larger AI pipeline built for Coding in Color (CIC) to analyze and improve student check-ins.

---

🎯 Purpose

Students are required to submit daily check-ins describing their work. However, messages can vary widely and may include:

• Announcements
• Questions
• Off-topic messages

This model filters those inputs and ensures only valid check-ins are processed further.

---

🧠 Model Details

• Task: Binary Text Classification
• Labels:
  • CHECKIN → Valid check-in
  • NOT_CHECKIN → Not a check-in
• Architecture: Transformer-based (fine-tuned)

---

📦 Use Cases

• Pre-processing filter for check-in pipelines
• Slack/Discord message classification
• Productivity tracking systems
• Educational workflow automation

---

🧪 Example

Input

Today I worked on fixing my API deployment issue and tested endpoints

Output

{
  "label": "CHECKIN",
  "confidence": 0.98
}

---

Pipeline Integration

This model is the first step in a multi-model pipeline:

• Check-in Detection (this model)
• Check-in Quality Classification
• Roadblock Detection
• Check-in Generation (if needed)

---

Usage (Python)

from transformers import pipeline

classifier = pipeline("text-classification", model="mjpsm/check_in_or_not_v2")

result = classifier("Today I worked on building an API")
print(result)

---

Future Improvements

• Improve edge case detection (short messages)
• Add multi-label classification (e.g., check-in + question)
• Expand dataset diversity

---

👤 Author

• Mazamesso Meba
• Information Systems Student | AI Developer