phoca-vitulina_yolo11x-seg

YOLO11x Phoca vitulina Segmentation Model (2025)

This is a specially trained YOLO model (based on the Ultralytics YOLO architecture) for Phoca vitulina segmentation and detection, trained based on the North Pacific Wildlife Consulting LLC dataset collected from drone imagery.

🔒 The model was trained exclusively on proprietary data and is provided for inference only.

📊 Model Details

• Task: Instance segmentation
• Class: Phoca vitulina (common seal)
• Input size: 640×640 (auto-resized)
• Format: PyTorch .pt (Ultralytics YOLO11x-seg)
• Trained on: Drone imagery dataset by North Pacific Wildlife Consulting LLC

🚀 How to Use

You can load and run inference directly from Hugging Face using the ultralytics library:

Manual Download

You can also download the model file directly:

📥phoca-vitulina_yolo11x-seg.pt

Automatic download via Hugging Face Hub

from huggingface_hub import hf_hub_download
import shutil
import os

from ultralytics import YOLO

filename="phoca-vitulina_yolo11x-seg.pt"
# Download to the cache
cached_path = hf_hub_download(
    repo_id="NPWCLLC/phoca-vitulina_yolo11x-seg",
    filename=filename,
)

# Copy to the desired folder
target_dir = "./models"
os.makedirs(target_dir, exist_ok=True)
target_path = os.path.join(target_dir, filename)

shutil.copy(cached_path, target_path)
print(f"✅ Model copied to: {target_path}")

# Making a prediction
model = YOLO(target_path)

results = model("images/image1.jpg")
results[0].save("output.jpg")
print("Done! Result saved to output.jpg")