Request for task-level evaluation aggregation protocol/code for Tables 1 and 2

[Jeremy2024]

Thank you for releasing the LLaRS paper and codebase. Your work has been immensely helpful to me.

The repository provides implementations for PSNR, SSIM, SAM, and ERGAS, and the test pipeline appears to log metrics separately for each dataset_name. However, I could not find the task-level evaluation procedure used to obtain the CR, PS, SR, HR, STF, DN, DB, DS, LS, EQ, and BE results in Tables 1 and 2.

In particular, could you please clarify or release code/configuration for:

1. Which dataset test splits are included in each task-level result;
2. How metrics from multiple underlying datasets within a task are aggregated;
3. Any task-specific preprocessing, channel selection, normalization, or other evaluation settings;
4. A script/config that reproduces the task-level columns and the final Avg. values in Tables 1 and 2.

Thank you again for your time and for sharing the project.