Submissions will be in the form of a .csv file. The first column should contain a list of all the image names in the test set. The remaining columns should contain your model's top five predicted classes with the most confident in the left and the least confident in the right. The columns should be named exactly what they are in the sample_submission.csv headers. 

## Example Submission File:
    Image Name,Class1,Class2,Class3,Class4,Class5
    "ccc83541.jpg",giovanni-boldini,leonardo-da-vinci,caravaggio,fyodor-vasilyev,george-morland
    "5adf0be9.jpg",andrea-del-sarto,giorgio-vasari,thomas-gainsborough,louise-elisabeth-vigee-le-brun,ivan-vladimirov
    "28d1c9fd.jpg",ivan-kramskoy,paolo-uccello,alfred-stevens,andrea-del-sarto,william-hogarth

Your submissions will be evaluated using a custom scoring scheme that balances accuracy with fair representation of all artists. For each painting in the test set, your model's prediction will receive a score based on its accuracy: 1 point for correct prediction (top 1), 1/2 point for top 2, 1/3 point for top 3, 1/4 point for top 4, 1/5 point for top 5, and 0 points if not in the top 5. An average score will be calculated for each artist across their paintings. The final score will be the average of all these artist-specific scores, ensuring that each artist contributes equally to the total score, regardless of how many of their works are in the test set. This method incentivizes models to perform well across all artists, not just the most common ones.

Approximately half of the training instances will be made private. The leaderboard will only evaluate on public data during the competition. The final rankings will be determined based on performance on the full dataset, including both public and private instances. This makes things a little more interesting as you won't be sure who truly is winning.