Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Create high-level selector API #124

Open
micahjsmith opened this issue Feb 22, 2019 · 0 comments
Open

Create high-level selector API #124

micahjsmith opened this issue Feb 22, 2019 · 0 comments
Assignees

Comments

@micahjsmith
Copy link
Contributor

The tuner provides a high-level propose/add API, in which it maintains an internal state of parameters that have been tried and the associated score, and updates its state in response to these user commands. At a low-level, it uses fit/predict to update its own meta-model.

The selector, in contrast, only provides a low-level API, in which it does not maintain any internal state and requires the user to pass the history of all selections to select.

We should create a high-level selector API, such as "select/add" or "select/record" or "select/report" in which the selector maintains the record of past trials. Then, the current select method can be replaced by the high-level select method, with compute_rewards and bandit already serving as a low-level API.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants