Submission Question

[Somesh Saini]

Good evening,

Hello. I had a few questions about the submission process and what exactly is expected to be submitted. 

1. Are we submitting a pre-trained model? I saw that this was an option in last year's challenge.

2. If not, how should I format the code to prepare the different datasets? Or will the datasets already be able to be extracted into the .hea and .dat format? If it's necessary, how do I make my model follow the folder paths of the output of the datasets?

3. The model would input all the ecg data from all three datasets and be trained off of that, correct?

4. Will the Chagas labels.csv be available to train the model on? Should the model look at the .hea file and read the Chagas labels off that?

Please let me know if I need to clarify any of my questions.

Best regards,

Somesh Saini

[PhysioNet Challenge]

Dear Somesh,

Please see answers to your questions below.

1. Are we submitting a pre-trained model? I saw that this was an option in last year's challenge.

Yes, please submit both working training code and a pre-trained model (unless you cannot fit a pre-trained model in your repository). For each entry, we will run your training code on the training set and evaluate the resulting model on the validation set, but the pre-trained model is helpful as well. (Please see previous questions about pre-trained models in the Challenge forum for more information.)

2. If not, how should I format the code to prepare the different datasets? Or will the datasets already be able to be extracted into the .hea and .dat format? If it's necessary, how do I make my model follow the folder paths of the output of the datasets?

You do not need to prepare the data. We will prepare the data and run your code using the commands in this README. In particular, we will prepare the data using the commands in the data preparation section of the README:
https://github.com/physionetchallenges/python-example-2025?tab=readme-ov-file#how-do-i-create-data-for-these-scripts

3. The model would input all the ecg data from all three datasets and be trained off of that, correct?

You can decide which data to use to train your model and how to use the data to train your model. You may or may not want to treat data from different sources differently.

4. Will the Chagas labels.csv be available to train the model on? Should the model look at the .hea file and read the Chagas labels off that?

Given the differences in data variables and formats between the different data sources, the inputs to the data preparation scripts will not be available. Your code should read the WFDB files for the data and labels. Please check the example code for file I/O functions and the data generated by the data preparation scripts in these repositories for the data that we will use with your code:
https://github.com/physionetchallenges/python-example-2025
https://github.com/physionetchallenges/matlab-example-2025

The best way to prepare a working approach is to use the above example code as a template and to submit your code for evaluation (after trying it locally) to see if it works. The end of the unofficial phase is approaching quickly, so please do not wait until the final days or hours so that we have time to provide feedback and you have time to make any updates before the deadline!

Best,
Matt
(On behalf of the Challenge team.)

Please post questions and comments in the forum. However, if your question reveals information about your entry, then please email info at physionetchallenge.org. We may post parts of our reply publicly if we feel that all Challengers should benefit from it. We will not answer emails about the Challenge to any other address. This email is maintained by a group. Please do not email us individually.