Failed to run
I simulated all the exact constraints as mentioned and ran it locally, it gave me an output. When I uploaded the file, its giving failed to run. What could be the cause of this?
2 Replies
suppressedmama24d agoHi flyingfreak,
will take a look and revert to you.
Best
Marc
suppressedmama24d agoHi flyingfreak — thanks for submitting. Your submission was rejected at the quick-validation stage, but the cause is a small, fixable one, not a problem with your modeling logic.
The short version: predict() must return a wide DataFrame of shape (n_timestamps, n_tickers) — one row per timestamp, one column per ticker — exactly the same layout as the target you get in train(). If predict() returns a long/stacked Series instead, the validator's cross-sectional de-mean check (which operates across tickers, i.e. across columns) can't run and the submission errors out before it's ever scored.
A good way to catch this before submitting: run python runner.py your_submission.py locally. It uses the exact same validation path as our server, so anything that fails there will fail on submission too — and anything that passes locally will run here.
I've emailed you the specifics and a one-line fix. Once you adjust the return shape you should be good to resubmit. 👍
Best
Marc
Sign in to reply.