This project is a Classification model which classifies the images for Handwritten digits.
The dataset used for trang the model is MNIST Handwritten Digits Dataset.
Inorder to evaluate the model on your own dataset:
- Make your dataset in the same format as of MNIST training dataset.
- Then run the ocr.sh file
- When it asks for path of dataset, enter the full path to it. Example:
$ ./ocr.sh
Enter path to the dataset:[YOUR FILE PATH]
If your dataset is a collection of images then:
- Run the ocr1.sh file
- When it asks for path of dataset, enter the full path to it. Example:
$ ./ocr.sh
Enter path to the dataset:[YOUR FILE PATH]
- It will make a file named output.txt with all the predicted labels.