metadata

license: agpl-3.0
language:
  - en
datasets:
  - nyuuzyou/planespotting
base_model:
  - Ultralytics/YOLO11
library_name: ultralytics
pipeline_tag: object-detection

Model Card for Planespotting-YOLO11

Model Summary

This dataset contains the results of fine-tuning YOLO11 models (n, s, m, l, x variants) on the Planespotting Dataset for aircraft detection, specifically for detecting aircraft in amateur photography taken with ordinary smartphones or low-quality airport cameras. The model was trained to detect aircraft on approach (often appearing as small dots in the sky) as well as aircraft at closer ranges.

Model Details

Model Variants

The fine-tuning was performed on multiple YOLO11 variants:

YOLO11n: Nano variant (smallest, fastest)
YOLO11s: Small variant'
YOLO11m: Medium variant
YOLO11l: Large variant
YOLO11x: Extra large variant (largest, most accurate)

Training Configuration

Image Size: 640 pixels
Epochs: 100
Training Split: 9,885 images
Validation Split: 93 images
Object Classes: 1 (aircraft)

Weights

The fine-tuned model weights are provided in the weights directory:

best.pt: Best model weights according to validation metrics
last.pt: Final model weights after completing all training epochs