Review : All issues in one
- General: remove ./data mention from everywhere.
- Evaluation criteria: remove log loss, remove the 2 invalid equation error in this section. log loss cannot be used as in submission the probability is not being taken , only binary values are taken.
- Submission:
small change in the first sentence.
Prepare a csv containing header as
label
and predicted value as digit1
if bank notes are genuine and digit0
if the notes are forged , name the file assubmission.csv
. - Files: i am not writing this explicitly. follow the pattern of https://gitlab.aicrowd.com/aicrowd/practice-challenges/aicrowd_ADCLK_challenge, the takeaway is to NOT use words like validation and training here, we do not want any prejudice for the solver. give information as clearly as possible.
- Dataset: The basic idea for dataset is to be very detailed and simple to understand. your dataset is good in this aspects. next issue highights small mistakes in dataset
Edited by shubhankar.sb