Scribd
Upload
6 views1 page

Deep Learning - Assignment

This assignment focuses on image classification using deep learning with the `Imagenette` dataset. It consists of three parts: building a basic CNN, implementing regularization techniques, and applying transfer learning with a pre-trained model on the CIFAR10 dataset. Students are required to report on their model architectures, training results, and comparisons between different approaches.

Uploaded by

Pritesh Das
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
6 views1 page

Deep Learning - Assignment

This assignment focuses on image classification using deep learning with the `Imagenette` dataset. It consists of three parts: building a basic CNN, implementing regularization techniques, and applying transfer learning with a pre-trained model on the CIFAR10 dataset. Students are required to report on their model architectures, training results, and comparisons between different approaches.

Uploaded by

Pritesh Das
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Deep Learning _Assignment

This assignment covers image classification with deep learning. You will train and evaluate two
different categories of networks on the `Imagenette` dataset. For each model, it is recommended
that you implement them using [PyTorch Lightning].

I. A Basic CNN

The first network will be a basic CNN. This network should include some number of convolutional
layers followed by fully connected layers. There is no size requirement for this network nor is there
a performance requirement. Train the model until convergence. Implement some form of early
stopping in case the model begins to overfit.
In your report, describe the chosen architecture and include the training loss, validation loss, and
final test accuracy of the model.

II. Regularization
Add regularization in the form of data augmentation or dropout. Train the model until convergence.
In your report, describe your choice of data augmentation and provide a clear comparison of the
model with and without regularization.

III. Transfer Learning

Transfer learning is an effective way to leverage features learned from another task into a new task.
For this part, use a model that was trained on the `Imagenette` dataset and fine-tune it is using the
CIFAR10 dataset.
Using a model from a previous run, re-train it from scratch on the CIFAR10 dataset. Take the same
model and initialize it with pre-trained weights from the Imagenette dataset. With the pre-trained
model, fine-tune it on the CIFAR10 dataset.
In your report, describe the pre-trained model you chose to use and include the fine-tuning training
plots along with the final model accuracy

You might also like