The Chinese MNIST image classifcation is a machine learning model that I've created that is able to classify fifteen chinese characters. The data was collected from https://www.kaggle.com/datasets/gpreda/chinese-mnist. The model was created on Tensorflow using transfer learning (the convolutional base uses a pretrained VGG16 where all except the last conv2d block is frozen). Finally, the model was deployed on a website using Flask.
A source of inspiration for this project: https://devpost.com/software/text-to-text
List of files/directories:
-Models: include all the deep learning models I've trained
-app.py: the app itself
-chinese_mnist.ipynb: the jupyter notebook used to process the data, train the model, and save the model
-model.py: a python script that initializes the model
-user_interface.html: the website