Recover best cell from all random trainings

capstone2019-neuralsearch / AC297r_2019_NAS

Harvard IACS Data Science Capstone: Neural Architecture Search (NAS) with Google

5 stars 1 forks source link

Closed jlaasri closed 4 years ago

jlaasri commented 4 years ago

This requires scraping the random search directories to find which cell got the best validation error and see how it performs on the test set.

jlaasri commented 4 years ago

Done with random_search_comparison*.py