Submissions will be evaluated based on  Mean Squared Error MSE, which measures the average magnitude of the errors between the predicted and actual values of the hidden numeric features. The model with the lowest MSE will be declared the winner.

The goal is to minimize this error, ensuring the model makes accurate predictions for the hidden molecular features based on the SMILES string input.

Lookup for the submition sample file to know how the to prepare your submition file.