A dataset consists of 12000 images split into two parts - 10000 train and 2000 validation samples. The images belong to 10 classes. You can download the dataset here.
The classes being
- Amphibia
- Arachnida
- Fungi
- Mammalia
- Plantae
- Animalia
- Aves
- Insecta
- Mollusca
- Reptilia
The task is to train a Convolutional Neural Network to predict which class the object in each image belongs to. You may make use of the sample code provided to start with.
-
You have to work individually.
-
You are requested to implement and train your model from scratch and report the accuracy on the validation split of the dataset.
-
If you are done with your predictions, you are requested to submit the Jupyter Notebook on this form.