NSL YOLOv8m + EfficientNet-B0 Baseline

This repository contains trained checkpoints and evaluation artifacts for the Nigerian Sign Language fixed YOLO + CNN experiment.

Results

YOLOv8m detection:

  • Test mAP50: 91.23%
  • Test mAP50-95: 60.59%
  • Test precision: 88.90%
  • Test recall: 83.51%

EfficientNet-B0 classification:

  • Test top-1: 90.59%
  • Test top-5: 96.81%
  • Test macro-F1: 89.19%

Files

  • yolov8m_nsl_fixed_best.pt: best YOLOv8m detector
  • cnn_efficientnet_b0_best.pt: best EfficientNet-B0 classifier checkpoint
  • summary.json: metrics and experiment metadata
  • class_names.json: class index mapping
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support