Loading
Have a ComfyUI workflow? Get a readable Workflow Doc in your browser.Make a workflow readable.Try it now ↗
YOLOv11x fine-tuned for drone detection on a combined dataset of 17,351 images from two public Roboflow datasets. Extends doguilmak/Drone-Detection-YOLOv11x with broader training data for improved generalisation on...
Model source
Source description
YOLOv11x fine-tuned for drone detection on a combined dataset of 17,351 images from two
public Roboflow datasets. Extends doguilmak/Drone-Detection-YOLOv11x
with broader training data for improved generalisation on diverse backgrounds and ground-camera perspectives.
from ultralytics import YOLO
model = YOLO("weights/best.pt")
results = model("your_image.jpg", conf=0.25)
results[0].show()
Sources
1 sourceVerified Sep 16
Model artifacts
1 artifactSource excerpts
2 excerptsFor ONNX inference:
import onnxruntime as ort
import numpy as np
# Input: [1, 3, 640, 640] float32 RGB, letterboxed
# Output: [1, 300, 6] — NMS baked in, format [x1, y1, x2, y2, score, class]
session = ort.InferenceSession("weights/best.onnx")
| Property | Value |
|---|---|
| Architecture | YOLOv11x (Ultralytics) |
| Input size | 640 × 640 |
| Classes | 1 — drone |
| Parameters | 56.8M |
| GFLOPs | 194.4 |
| Base checkpoint | doguilmak/Drone-Detection-YOLOv11x |
| License | MIT |
| Parameter | Value |
|---|---|
| Epochs | 16 (early-stopped, patience=10) |
| Batch size | 16 |
| Image size | 640 |
| Optimizer | default (AdamW) |
| Hardware | NVIDIA RTX PRO 4500 Blackwell (32 GB) |
Training data — combined from two Roboflow datasets:
| Dataset | Images (train) | Source |
|---|---|---|
| Zhejiang University Drone Detection | 10,192 | Ground-camera perspective |
| DetFly (YOLOv8 DetFly-02) | 4,839 | Varied backgrounds incl. sky clutter |
All labels were unified to a single drone class (index 0).
Evaluated on the combined validation set (2,320 images, conf=0.25, IoU=0.5):
| Model | Precision | Recall | [email protected] | [email protected]:0.95 |
|---|---|---|---|---|
| This model | 0.775 | 0.606 | 0.555 | 0.279 |
| Base v11x (on its own val set, ~350 imgs) | 0.922 | 0.831 | 0.905 | 0.546 |
The base model's higher numbers are on a smaller, in-distribution validation set. This model was evaluated on a harder, more diverse out-of-distribution set — a better proxy for real-world performance across varied backgrounds.
Real-time drone detection from fixed or mobile cameras. Designed for scenarios with varied backgrounds (sky, treeline, urban). Not validated for maritime or night-vision use.
--- license: mit base_model: doguilmak/Drone-Detection-YOLOv11x tags: - object-detection - drone-detection - yolov11 - ultralytics - onnx datasets: - zhejiang-university-china-dliq1/drones-detection-with-yolov8 - yolov8-drone-detection/yolov8_detfly-02 --- # YOLOv11x Drone Detector — Fine-tuned YOLOv11x fine-tuned for drone detection on a combined dataset of 17,351 images from two public Roboflow datasets. Extends [`doguilmak/Drone-Detection-YOLOv11x`](https://huggingface.co/doguilmak/Drone-Detection-YOLOv11x) with broader training data for improved generalisation on diverse backgrounds and ground-camera perspectives. ## Usage ```python from ultralytics import YOLO model = YOLO("weights/best.pt") results = model("your_image.jpg", conf=0.25) results[0].show() ``` For ONNX inference: ```python import onnxruntime as ort import numpy as np # Input: [1, 3, 640, 640] float32 RGB, letterboxed # Output: [1, 300, 6] — NMS baked in, format [x1, y1, x2, y2, score, class] session = ort.InferenceSession("weights/best.onnx") ``` ## Model details | Property | Value | |---|---| | Architecture | YOLOv11x (Ultralytics) | | Input size | 640 × 640 | | Classes | 1 — `drone` | | Parameters | 56.8M | | GFLOPs | 194.4 | | Base checkpoint | `doguilmak/Drone-Detection-YOLOv11x` | | License | MIT | ## Training | Parameter | Value | |---|---| | Epochs | 16 (early-stopped, patience=10) | | Batch size | 16 | | Image size | 640 | | Optimizer | default (AdamW) | | Hardware | NVIDIA RTX PRO 4500 Blackwell (32 GB) | **Training data — combined from two Roboflow datasets:** | Dataset | Images (train) | Source | |---|---|---| | Zhejiang University Drone Detection | 10,192 | Ground-camera perspective | | DetFly (YOLOv8 DetFly-02) | 4,839 | Varied backgrounds incl. sky clutter | All labels were unified to a single `drone` class (index 0). ## Evaluation Evaluated on the combined validation set (2,320 images, conf=0.25, IoU=0.5): | Model | Precision | Recall | [email protected] | [email protected]:0.95 | |---|---|---|---|---| | **This model** | **0.775** | **0.606** | **0.555** | **0.279** | | Base v11x (on its own val set, ~350 imgs) | 0.922 | 0.831 | 0.905 | 0.546 | The base model's higher numbers are on a smaller, in-distribution validation set. This model was evaluated on a harder, more diverse out-of-distribution set — a better proxy for real-world performance across varied background...
Source context: 68 downloads · 2 likes · Pipeline object-detection · Library ultralytics · Repo FilippTrigub/yolov11x-drone-finetuned