Identifying dog breeds can be difficult due to the high similarity between breeds and the number of possible classes. In this study, we aim to develop a deep learning model using a convolutional neural network (CNN) to classify dog breeds. We used a dataset of over 20,000 labeled images from 120 distinct breeds. We analyzed the impact of model complexity, regularization techniques, and learning rate of our custom CNN model's performance. We also explored the use of fine-tuning a pre-trained model like ResNet50V2 for classification. Our best-performing model achieved a 66.92% accuracy, highlighting the potential of CNNs in dog breed classification.

Report - Project Report -Predicting Pooches.pdf

This project aims to classify dog breeds using convolutional neural networks. The code is contained in a Jupyter Notebook.

• A Google account (for accessing Google Colab) or other Jupyter notbook platform

1. Open Google Colab (https://colab.research.google.com/) and sign in with your Google account.

2. In Google Colab, click on File in the menu and select Open notebook.

3. Open the Predicting Pooches.ipynb notebook file.

1. In the notbooke, you can specifiy the modle save path before running.

2. To run the entire notebook, click Runtime in the menu, then Run all. Alternatively, you can run each cell individually by clicking on a cell and pressing Shift + Enter.

• If you run the entire notbook it will train each model, THIS IS A 12+ HOUR TASK SO PROBABLY DONT WANT TO DO THIS

3. The notebook will execute the code in each cell and display the output below the respective cell.

4. You can modify the code in the cells and re-run them to experiment with different configurations or settings.

Please refer to the comments and markdown cells within the notebook for more information about the code and its functionality.