You can deploy this model and start earning money today!

Sph3inxz/
eaglenet-large-specialist-r1

Name: eaglenet-large-specialist-r1
Rating: 2.5 (6 reviews)

Views

No views yet

EAGLE-Net large-drone specialist (`large-specialist-r1`)

Vanilla YOLO11s detector for the EAGLE-Net hybrid pipeline. Complements the NWD+P2 small-object specialist (P1) for drones larger than ~50 px.

Trained on filtered SynDroneVision large-drone boxes (see scripts/filter_large_drones.py and eaglenet train large).

Usage

python

from ultralytics import YOLO

model = YOLO("hf://Sph3inxz/eaglenet-large-specialist-r1/best.pt")
results = model.predict("video.mp4", imgsz=640, conf=0.30)

Hybrid inference (with P1 small specialist):

bash

eaglenet infer demo \
  --weights <p1-best.pt> \
  --weights-large hf://Sph3inxz/eaglenet-large-specialist-r1/best.pt \
  --imgsz-large 640 --conf-large 0.30 \
  --source clip.mp4 --out tracked.mp4

Files

File	Description
`best.pt`	Ultralytics checkpoint (YOLO11s, 1-class drone)

Training (summary)

Data: SynDroneVision (large boxes only, filtered from full corpus)
Base: yolo11s.pt (COCO), stock CIoU, no P2 head
imgsz=640, scale aug 0.5, single class drone
Intended for large boxes; poor Anti-UAV v4 scores are expected (different regime from P1)

Source

EAGLE-Net

Sph3inxz/eaglenet-large-specialist-r1

EAGLE-Net large-drone specialist (large-specialist-r1)