Pull Request: Enhancements in Dataset Inclusion Flexibility
Overview
This pull request introduces key enhancements in the flexibility of dataset inclusion. It significantly simplifies the process of dataset aggregation and addresses two pivotal issues identified in our GitHub repository.
Enhancements
Updated README.md to accurately reflect the new changes and guide users through the updated processes.
Refactored aggregate_data.py script to correct the filename for splits_final.json and enable more flexibility. Now, users can specify which datasets to include in the aggregated dataset, and customize the name, description, and number of stratification folds in final_splits.json.
Addressed GitHub Issues
Issue #5: Resolved the filename inconsistency with splits_final.json. The script autonomously places the file in the designated directory, specifically nnUNet_preprocessed/DATASETXXX_NAME, eliminating the need for a separate command to specify the file's location.
Issue #6: The script modifications facilitate an ablation study on data aggregation. It allows for a comparative analysis between models trained on aggregated Bright Field (BF) datasets and dedicated BF models.
Conclusion
These changes enhance the project's functionality and user experience, addressing community-reported issues and improving data handling capabilities. Feedback and further contributions are welcome to refine these features.
Pull Request: Enhancements in Dataset Inclusion Flexibility
Overview
This pull request introduces key enhancements in the flexibility of dataset inclusion. It significantly simplifies the process of dataset aggregation and addresses two pivotal issues identified in our GitHub repository.
Enhancements
README.md
to accurately reflect the new changes and guide users through the updated processes.aggregate_data.py
script to correct the filename forsplits_final.json
and enable more flexibility. Now, users can specify which datasets to include in the aggregated dataset, and customize the name, description, and number of stratification folds infinal_splits.json
.Addressed GitHub Issues
splits_final.json
. The script autonomously places the file in the designated directory, specificallynnUNet_preprocessed/DATASETXXX_NAME
, eliminating the need for a separate command to specify the file's location.Conclusion
These changes enhance the project's functionality and user experience, addressing community-reported issues and improving data handling capabilities. Feedback and further contributions are welcome to refine these features.