:technologist: Implement test_step to generate submission CSV file

What I am changing

[x] Convert the predicted masks on the validation set to binary run length encoding format, and format it into a dataframe table.
[x] Documented how to create the output CSV file with a one-liner python trainer.py test command
[x] Added unit tests to ensure that the implementation in ChaBuDNet.test_step works

Run python trainer.py test --model.submission_filepath=submission.csv in the command-line

Resources used to create the scripts in this PR:

Sample submission.csv file - https://huggingface.co/datasets/chabud-team/chabud-ecml-pkdd2023/blob/main/sample_submission.csv
Official script to generate CSV file - https://huggingface.co/datasets/chabud-team/chabud-ecml-pkdd2023/blob/main/create_sample_submission.py
Note at https://huggingface.co/spaces/competitions/ChaBuD-ECML-PKDD2023 under the Dataset -> Evaluation -> How to submit

The submissions must be in csv format with three columns: id, index, rle_mask. Remember to “explode” index and rle_mask, having a single value for each line.
Documentation about Binary Run Length Encoding (BRLE) format - https://trimsh.org/trimesh.voxel.runlength.html#trimesh.voxel.runlength.dense_to_brle