Vaibhavsh0120/ATM-Theft-Detection

ATM Theft Detection YOLOv8 Model

This folder is intended to be pushed as a standalone Hugging Face model repository.

Model Summary

The model detects two classes for ATM monitoring:

• Face_Covered
• Face_Uncovered

Files

• weights/best.pt: published YOLOv8 checkpoint
• exports/best_full_integer_quant.tflite: optional quantized export copied from local training

Local Usage

from ultralytics import YOLO

model = YOLO("weights/best.pt")
results = model("example.jpg")

Source Workflow

The canonical local training flow lives in the root repository under training/. After training, push the latest model and Space snapshots with:

python scripts/push_hf.py