Downloads · 30 days
0
power-equipment-fault-detector
aj3944/power-equipment-fault-detector
power-equipment-fault-detector is a machine learning model from aj3944. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as mit.
This model detects and classifies high-voltage electrical transmission line faults and power equipment defects with 94.9% Top-1 Accuracy, trained on the Power Equipment Image Dataset using an NVIDIA RTX 5090 GPU.
Access
Public
Updated Sep 3, 2026
Repo size
8.6 MB
Likes
1
Public
Repo makeup
Click a slice to open those files.
.pt8.6 MB · 100%
At a glance
- License
- mit
- Access
- Public
- Created
- Sep 3, 2026
- Updated
- Sep 3, 2026
- SHA
- 5ff96f2a
Model card
From the Hugging Face model README
Power Equipment & Transmission Line Fault Detector
This model detects and classifies high-voltage electrical transmission line faults and power equipment defects with 94.9% Top-1 Accuracy, trained on the Power Equipment Image Dataset using an NVIDIA RTX 5090 GPU.
Target Defect Classes
| Class Code | Defect Name | Severity | Failure Physics & Corrective Action |
|---|---|---|---|
| czjyz_ps | Porcelain Post Insulator Breakage | CRITICAL | Loss of creepage distance; urgent emergency replacement to prevent arc flashover. |
| czjyz_sl | Porcelain Post Insulator Flashover | CRITICAL | Conductive carbonization track; high-pressure cleaning or replacement. |
| czjyz_lw | Porcelain Post Insulator Crack | HIGH | Mechanical fatigue fracture under wind vibration; schedule maintenance replacement. |
| dx_jypps | Conductor Wire Insulation Damage | HIGH | Core exposure; install repair armor sleeve over conductor. |
Performance Metrics
- Top-1 Accuracy: 94.9% (on 2,002 validation images)
- Top-5 Accuracy: 100%
- Inference Speed: ~0.3 ms / image on GPU (14 ms HTTP endpoint)
- Model Size: 3.2 MB (YOLO11n-cls)
Interactive Jupyter Notebook
Download or view the included notebook power_equipment_fault_detection.ipynb to inspect dataset loading, training curves, and visualization.
Production Hugging Face Space
Experience live predictions via the hosted Space: Power Equipment Fault Inspector
Quick Usage
- License
- mit
- Created
- Sep 3, 2026
- Updated
- Sep 3, 2026