Chili Grade Classification (Python)

This repo contains implemetation of Chili Grade Classification built using Python

Follow Fastai_ChiliGrade_classifier.ipynb file for training and inference purpose. Requirements of the project is also mentioned in given jupyter notebook.

Training details

• Batch-size : 4
• GPU provide : Colab
• GPU model : Tesla t4
• Epochs : 430

Model is overfitted as there is not enough dataset to make a generaised dataset but accuracy is 100 percent on 5% validation set.

Validation set Confusion matrix (5% of dataset)

Confusion matrix on 50% of complete dataset

Notes

To prevent underfiting we can add augumentations of different kind to balance the dataset and to generate more samples.