MLCommons Algorithmic Efficiency is a benchmark and competition measuring neural network training speedups due to algorithmic improvements in both training algorithms and models.
The final hparam point for the external prize_qualification_baselines has dropout set to 0.1. To repro either run the last point with hparam_start_index=4:
Description
FastMRI returns the following error:
Logs: https://gist.github.com/priyakasimbeg/e2ecdd85340994fc15e01738dc7d8d2c
Steps to Reproduce
The final hparam point for the external prize_qualification_baselines has dropout set to 0.1. To repro either run the last point with
hparam_start_index=4
:Or remove the other hparam points in the
tuning_search_space.json
and set tuning trials to 1.