The Modified National Institute of Standards and Technology (MNIST) database contains images of handwritten digits. The MNIST database is popular for machine learning and deep learning exploration and study.
Training data The model-building-code that is generated by the flow editor for this tutorial expects training data to be in pickle file format, and divided into 3 files:
- mnist-keras-train.pkl : Images and labels for training the model
- mnist-keras-validate.pkl : Images and labels for calculating the loss and evaluating metrics at the end of each training iteration
- mnist-keras-test.pkl : Images and labels used to measure the performance of the trained model
Test data The .zip file available from here contains the three training data files already mentioned as well as an addition file:
- mnist-keras-test-payload.json
How to download the sample files Click the download icon on the toolbar, then save the sample files to your computer. From there, you can add them to a project for use in sample applications.