Content

Dataset Link - https://www.kaggle.com/lsind18/gemstones-images
This dataset contains 3,200+ images of different gemstones. The images are grouped into 87 classes which are already divided into train and test data. All images are in various sizes and are in .jpeg format.

Objectives

The objective is to predict to which class a particular gemstone belongs.

Tools Used

Steps involved in making the model

1. Importing our data.
   
2. Data Augmentation.
   
3. Model Building.
   
4. Flattening and Adding Dense Layers.
   
5. Using vgg19 Tranfer learning techniques.
   
6. Passing it into the gradio app
7. Concluding our analysis by testing the model with some random user input.
   

Result

The Model with the help of gradio is able to make a decent prediction in which class a particular gemstone belongs to.