The evaluation of proposed solutions is based on Accuracy. The submission file is expected to be a .csv composed of two columns: ID and TARGET. The TARGET column should contain the predicted class label (integer from 0 to 9). The order of rows is not important: the image ID is instead extremely important. Use the same IDs provided in the test .csv file and associate each such ID to the image path reported there (which contains the ID itself to avoid confusion). A misalignment between IDs and predictions can impair the evaluation of your solution: we suggest to use the same DataFrame contained in test.csv and to just append a column with your predictions (keep the ID,TARGET pairs matched) and to then remove the "path" column before saving the DataFrame to .csv for submission.

## Submission File
For each image ID in the test set, you must predict a class label (from 0 to 9) for the TARGET variable. The file should contain a header (ID, TARGET) and have the following format:

    ID,TARGET
    201,0
    234,7
    172,4
    etc.