Submissions are evaluated on the following criteria:

•	Primary Metric: The competition’s primary evaluation metric is the F1-Score, which balances precision and recall to measure your model’s effectiveness in identifying AI-generated and human-created images. It is calculated at the 0.5 threshold. 

## Submission Format

Participants must submit their predictions in the specified format to be eligible for evaluation. Please follow the guidelines below to ensure your submission is valid:

1. **Test file**: Provided in the Data section as `test.csv`
2. **File Format**: Submissions should be in a CSV file format.
3. **Columns**:
   - The file must contain only two columns:
     - **id**: A unique identifier for each test image.
     - **label**: The predicted class for each image (e.g., 0 for human-created and 1 for AI-generated).
   - Any additional columns or index columns should be removed before submission.

Example Format:

    id,label
    1,0
    2,1
    3,0
    4,1
    etc.

## Notebook Requirements

In addition to the prediction file:

- Submitted notebooks should be **clear, well-documented, and easily replicable**.
- Ensure that:
  - **Code is structured and annotated** to facilitate understanding.
  - **Results are reproducible** when the notebook is executed with the provided dataset.

Submissions that lack clarity or fail to meet reproducibility standards may not be considered for final evaluation.