svm-classification-cat-and-dog

Cat vs Dog SVM Classification Model

This repository contains the trained SVM model and preprocessing artifacts for cat vs dog image classification.

Model Files

• svm_best_model.pkl.gz - Compressed SVM model (RBF kernel, ~15MB)
• scaler.pkl - Feature scaling preprocessing object
• label_encoder.pkl - Label encoding for classes
• metadata.json - Model metadata and training information

Usage

from huggingface_hub import hf_hub_download
import joblib
import gzip
import pickle

# Download model files
model_path = hf_hub_download("your-username/cat-dog-svm-model", "svm_best_model.pkl.gz")
scaler_path = hf_hub_download("your-username/cat-dog-svm-model", "scaler.pkl")
encoder_path = hf_hub_download("your-username/cat-dog-svm-model", "label_encoder.pkl")

# Load model
with gzip.open(model_path, 'rb') as f:
    model = pickle.load(f)

# Load preprocessing objects
scaler = joblib.load(scaler_path)
label_encoder = joblib.load(encoder_path)

Model Architecture

• Algorithm: Support Vector Machine (SVM)
• Kernel: RBF (Radial Basis Function)
• Features: HOG + Color Histogram + LBP
• Input Size: 128x128 RGB images
• Classes: ['cat', 'dog']