m7kael/yachtmaster-boat-classifier

YachtMaster Boat Image Classifier

A Vision Transformer (ViT) model fine-tuned to classify boat images into 5 categories:

• cruising: Photos of boats underway on the water
• exterior: Static dock, marina, or anchored views
• interior: Cabin, salon, galley, stateroom, or internal living spaces
• layout: Floor plans, general arrangement drawings, deck plans, or schematic diagrams
• technical: Machinery spaces, engine compartments, bilge, or technical areas

Usage

from transformers import ViTImageProcessor, ViTForImageClassification
from PIL import Image

# Load model and processor
processor = ViTImageProcessor.from_pretrained("m7kael/yachtmaster-boat-classifier")
model = ViTForImageClassification.from_pretrained("m7kael/yachtmaster-boat-classifier")

# Load image
image = Image.open("boat.jpg")

# Process and predict
inputs = processor(image, return_tensors="pt")
outputs = model(**inputs)
logits = outputs.logits

# Get predictions
predicted_class_idx = logits.argmax(-1).item()
predicted_class = model.config.id2label[predicted_class_idx]
print(f"Predicted class: {predicted_class}")

API Usage

import requests

API_URL = "https://api-inference.huggingface.co/models/m7kael/yachtmaster-boat-classifier"
headers = {"Authorization": "Bearer YOUR_TOKEN"}

def query(filename):
    with open(filename, "rb") as f:
        data = f.read()
    response = requests.post(API_URL, headers=headers, data=data)
    return response.json()

# Example
result = query("boat.jpg")
print(result)

Performance

Based on test set evaluation:

• Accuracy: Check test_results.json
• F1 Score: Check test_results.json
• Precision/Recall: Check test_results.json per class

Training

Fine-tuned from google/vit-base-patch16-224-in21k on custom boat image dataset.

• Base model: Vision Transformer (ViT) with patch size 16x16
• Input size: 224x224 pixels
• Number of classes: 5
• Framework: Hugging Face Transformers + PyTorch

Model Details

• Model type: ViTForImageClassification
• Base model: google/vit-base-patch16-224-in21k
• Input preprocessing: ViTImageProcessor
• Classes: 5 boat image categories
• Format: Safetensors (optimized for inference)

License

This model is released under the Apache 2.0 license.