I noticed that my smoke test passes but whenever I submit I get “IDs for submission are not correct.”
What could possibly lead to this issue if both results follow the submission format?
Minor addition is the test_features.csv given by the competition?
@Ethan-Sinson This means that the IDs (uid column) do not match the submission_format.csv provided in the competition runtime for the full test set. There are a number of reasons this could be happening (eg. correctly ordered rows, missing some uids, etc).
I recommend checking out the benchmark blog post for some handy code to make sure your rows and columns are in the right order. E.g.:
# Make sure the row order is the same as in the submission format
parsed_preds = parsed_preds.loc[features.index]
@kwetstone I’ll check this out, but I also do want to know whether test_features.csv should be on the data download page?
@Ethan-Sinson This is a code execution challenge, so test features are kept confidential. Your code will be used to generate predictions on the test features in the code execution runtime. The submission format is shared, so you can see the expected uids in the full test set.