# Submission pass smoke test but fail normal submission

**URL:** <https://community.drivendata.org/t/submission-pass-smoke-test-but-fail-normal-submission/11413>\
**Category:** Children’s Speech Recognition Challenge\
**Created:** [March 14, 2026, 3:24am UTC](https://community.drivendata.org/t/submission-pass-smoke-test-but-fail-normal-submission/11413 "2026-03-14T03:24:09Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![jialuli](https://avatars.discourse-cdn.com/v4/letter/j/a87d85/32.png) [@jialuli](https://community.drivendata.org/u/jialuli)\
**Post date:** [March 14, 2026, 3:24am UTC](https://community.drivendata.org/t/submission-pass-smoke-test-but-fail-normal-submission/11413/1 "2026-03-14T03:24:09Z")

</div>

I passed the smoke test but failed the normal submission. How can I view the full log? The downloaded log seems to be truncated after a certain number of lines. Also, can we assume that the real test set uses the same metadata format as the smoke test?

---

<div class="post-metadata">

**Author:** ![cszc](https://avatars.discourse-cdn.com/v4/letter/c/c68b51/32.png) [@cszc](https://community.drivendata.org/u/cszc)\
**Post date:** [March 14, 2026, 10:17pm UTC](https://community.drivendata.org/t/submission-pass-smoke-test-but-fail-normal-submission/11413/2 "2026-03-14T22:17:16Z")

</div>

As noted on the [Code Submission Format Page](https://www.drivendata.org/competitions/308/childrens-word-asr/page/978/), log output is limited to 500 lines and 300 characters per line. Yes, you can assume the real test set uses the same metadata format as the smoke test.

---

<div class="post-metadata">

**Author:** ![uamo](https://avatars.discourse-cdn.com/v4/letter/u/f17d59/32.png) [@uamo](https://community.drivendata.org/u/uamo)\
**Post date:** [March 14, 2026, 11:50pm UTC](https://community.drivendata.org/t/submission-pass-smoke-test-but-fail-normal-submission/11413/3 "2026-03-14T23:50:01Z")

</div>

I had the same problem, then realized that the normal test set has some long audios which cause Out of Memory errors, so reduced the batch size.
