Dog Breeds Classifier (AlexNet Fine-tuned Model)

Model Description

This is a fine-tuned version of the AlexNet model, designed to classify images into one of 10 different dog breeds. The original AlexNet architecture was pre-trained on the ImageNet dataset, and this version has been specifically adapted to classify dog breeds based on a custom dataset containing images of various dogs.

Model Details

Architecture

Base Model: AlexNet
Final Layer: The final fully connected layer has been modified to classify 10 dog breeds.
- Input to the last fully connected layer: 4096 features
- Output: 10 classes (one for each breed)

Dataset

The dataset used for fine-tuning consists of images of 10 different dog breeds, organized into training, validation, and test sets.

Dog Breeds Included:

Beagle
Chihuahua
Corgi
Dalmation
Doberman
Golden Retriever
Maltese
Poodle
Shiba Inu
Siberian Husky

Data Format

Image Format: JPG
Resolution: The images were resized to 227x227 pixels to match AlexNet’s input requirements.
Dataset Structure:
- train/: Contains the training images of dog breeds.
- valid/: Contains the validation images of dog breeds.
- test/: Contains the test images of dog breeds.

Source:

The dataset was obtained from Kaggle. You can access it here.

Training

The AlexNet model was fine-tuned using the following setup:

Optimizer: SGD
Learning Rate: 0.001
Momentum: 0.9
Loss Function: CrossEntropyLoss
Training Epochs: 10
Batch Size: 32

Usage

You can load and use this model for inference by following the code snippet below.

# Install Required Libraries
pip install torch torchvision huggingface_hub

import torch
import torch.nn as nn
from torchvision import models
from huggingface_hub import hf_hub_download

# Load the fine-tuned AlexNet model
model = models.alexnet(pretrained=False)
num_features = model.classifier[6].in_features
model.classifier[6] = nn.Linear(num_features, 10)

# Download model weights from Hugging Face Hub
model_path = hf_hub_download(repo_id="pramudyalyza/dog-breeds-alexnet", filename="alexnet_model.bin")

# Load the model state
model.load_state_dict(torch.load(model_path))
model.eval()

# Example inference
from PIL import Image
from torchvision import transforms

# Define transformation
transform = transforms.Compose([
    transforms.Resize((227, 227)),
    transforms.ToTensor(),
    transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
])

# Load and preprocess an example image
image = Image.open("path_to_image.jpg")
image = transform(image).unsqueeze(0)

# Perform inference
with torch.no_grad():
    output = model(image)
    predicted_class = output.argmax(dim=1)
    print(f"Predicted Dog Breed: {predicted_class.item()}")