https://github.com/maheera421/dog-breed-prediction-model

Multi-class image classification, where the goal is to predict different breeds of dogs using deep learning.
https://github.com/maheera421/dog-breed-prediction-model

boolean-array color-channel data-batch google-colab gpu ipython keras-tensorflow matplotlib-pyplot mobilenetv2 numpy-arrays os pandas-dataframe sklearn tensorflow2 tensorflowhub tensors

Last synced: about 1 month ago
JSON representation

Multi-class image classification, where the goal is to predict different breeds of dogs using deep learning.

• Host: GitHub
• URL: https://github.com/maheera421/dog-breed-prediction-model
• Owner: maheera421
• Created: 2025-01-24T15:15:54.000Z (3 months ago)
• Default Branch: main
• Last Pushed: 2025-02-03T16:57:44.000Z (3 months ago)
• Last Synced: 2025-02-03T17:43:14.421Z (3 months ago)
• Topics: boolean-array, color-channel, data-batch, google-colab, gpu, ipython, keras-tensorflow, matplotlib-pyplot, mobilenetv2, numpy-arrays, os, pandas-dataframe, sklearn, tensorflow2, tensorflowhub, tensors
• Language: Jupyter Notebook
• Homepage:
• Size: 5.86 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# 🐶 Dog Breed Prediction Model  

This project focuses on **multi-class image classification**, where the goal is to predict different breeds of dogs using deep learning. We leverage **TensorFlow and Transfer Learning** to classify images into multiple dog breeds with high accuracy.  

## 📌 Project Overview  

This deep learning project applies a **convolutional neural network (CNN)** to classify dog breeds. Using a **pretrained model from TensorFlow Hub**, we fine-tune the model to recognize and differentiate multiple breeds efficiently.  

The workflow follows the standard deep learning pipeline:  

1. **Getting the Data Ready** 🗂️  

   - Import and store dataset  

   - Load images and labels for training  

2. **Preparing the Data** 🔄  

   - Convert images into numerical tensors  

   - Split images and labels (X, y)  

   - Create a **custom validation set**  

   - Convert image tensors into **tensor tuples**  

   - Generate **data batches of size 32**  

3. **Choosing and Training the Model** 🎯  

   - Define input and output layers  

   - Use a **pretrained deep learning model from TensorFlow Hub** (transfer learning)  

   - Train the model using GPU for faster processing  

4. **Evaluating the Model** 📊  

   - Make predictions on new images  

   - Compare predictions with **ground truth labels**  

   - Analyze model performance using accuracy and loss metrics  

5. **Improving the Model** 🚀  

   - Experiment with different model architectures  

   - Fine-tune hyperparameters  

   - Optimize training strategies for better performance  

6. **Saving, Sharing, and Reloading the Model** 💾  

   - Save trained model for later use  

   - Reload model for future inference tasks  

## 📂 Dataset  

The dataset consists of images of **multiple dog breeds**, categorized into different classes. Each image is labeled with its respective breed, serving as the ground truth for training and evaluation.  

## 🖥️ Google Colab & GPU Usage  

This project was implemented using **Google Colab**, taking advantage of **GPU acceleration** for faster training. Running deep learning models on GPU significantly reduces training time compared to CPUs.  

## 🔧 Setup  

Key libraries used:  

- **TensorFlow 2.x**: For deep learning and model training  

- **TensorFlow Hub**: For pretrained models (transfer learning)  

- **Matplotlib & Seaborn**: For data visualization  

- **NumPy & Pandas**: For data manipulation  

## 📈 Model Performance  

The model is evaluated using standard metrics:  

- **Accuracy**: Measures the correct classification rate  

- **Loss**: Evaluates model errors