Unanswered
Hi! I Got A Question Regarding The Following Scenario: I Want To Try Out Different Hyperparmeter Settings Using Clear Ml, And I Want To Run The Experiments On My Local Computer. Is It Correct That This Requires A Git Repo?
so i figured out that we can use optimizer.start_locally() to do otherwise. But it seems that remotely (that is, using the optimizer.start() function that uses github apparently), I can only run 1 job at a time, while locally allows me to run as many experiments as I want specified by the 'max_number_of_concurrent_tasks' paramter. Is that normal?
122 Views
0
Answers
10 months ago
10 months ago