project-01-dl-hcr is the course project for EEL5840 19Fall Class at UF

Team Member:

• Wanyu Dong 👧 📧 wanyudong
• Zhiyun Ling 👦 📧 zhiyunl

• train_data.pkl - This is our train dataset, which is 6400 handwritten character images, including [a, b, c, d, h, i, j, k]. The size of these images around 50X50, but not fixed.
• finalLabelsTrain.npy - This is correspondent labels of train_data.plk.[1, 2, 3, 4, 5, 6, 7, 8]

• MyLoader.py - Functions to help loading dataset. Preprocess function included.

• train.py - Core file. Including the function of train_cnn. Main function is how to train the cnn model with train_data and train_labels, and important parameters set.
• test.py - Main function, test_cnn function and example of how to test your data.
• net_8_final.pth - Our trained cnn parameter set for both 2-class and 9-class, you can use this to test directly without training.

Use train.py to train, and test.py to test

• In train.py - the main function can run the training and no need for configuration.
• In test.py - the main function shows how to use test_cnn. Just modify pkl file names to load your test data, and get prediction on return of test_cnn.

#Have fun !