Skip to main content

Submit to Leaderboard

Local evaluation is not possible since model weights are encrypted and cannot be downloaded.
  1. Navigate to the Training section of your use case, choose an experiment, hover over the learning curves to identify the best performing validation cycle, then enter it into the “Submit for inference” field
  2. Click “Submit” to start inference on the test dataset
  3. View your model’s performance on the leaderboard tab once complete
Submit a Model Note: Track your team’s remaining daily submissions at the top of the use case page. The number of submissions is restricted to 5 to prevent overfitting on the test dataset.

Leaderboard Rankings

Navigate to the Leaderboard section of your use case (e.g., Breast Cancer Screening Use Case) to compare submitted models. Contributors are ranked by their best performing model score on the test data. Compare:
  • Final model score on test data
  • Model size
  • Energy consumption
  • Number of submissions
  • Remaining compute budget per team

Next Steps


Need Help?