The evaluation metric is a combination of accuracy and Mean Squared Error (MSE):  
$$
\text{Final Score} = 0.6 \cdot \text{Accuracy} + 0.2 \cdot( (1 - MSE) \cdot C_1 + (1 - MAE) \cdot C_2)
$$

where:  
- **Accuracy**: Fraction of correctly predicted categorical attributes (\(X_1, X_2, X_3\)).  
- **MSE**: Mean squared error for the speed (\(X_4\)) , between the predicted values and the ground truth values.  
- **MAE**: Mean absolute error for the video summary (\(X_5\)) , between the predicted values and the ground truth values.  
- **C**: A normalization constant to scale MSE/MAE to the range of accuracy.

**Public Leaderboard will display scores based on only 1/3rd of the data. Private Leaderboard with all all of the test data will be published after the competition ends.**