search

Neuraxio / Neuraxle

The world's cleanest AutoML library ✨ - Do hyperparameter tuning with the right pipeline abstractions to write clean deep learning production pipelines. Let your pipeline steps have hyperparameter spaces. Design steps in your pipeline like components. Compatible with Scikit-Learn, TensorFlow, and most other libraries, frameworks and MLOps environments.
https://www.neuraxle.org/
Apache License 2.0
608 stars 62 forks source link

Feature: Multi-Head AutoML optimization #473

Closed guillaume-chevalier closed 8 months ago

guillaume-chevalier commented 3 years ago

Is your feature request related to a problem? Please describe. https://stackoverflow.com/questions/67035671/is-it-possible-to-combine-multiple-pipeline-into-single-estimator-in-neuraxle-or

Describe the solution you'd like Ability to separate heads' optimization buy either running different AutoML loops within one pipeline, or else by allowing the optimization of multiple heads within one AutoML run using various metrics

Describe alternatives you've considered Separating pipelines and assembling them afterwards

Additional context See Stack Overflow thread

guillaume-chevalier commented 3 years ago

Idea: have everything we need to store be in the Hyperparams Repo, but by changing the evaluation metrics on the selected head (or output) we could do things properly. Thus, it'd be easy to reuse trials across various runs.

guillaume-chevalier commented 2 years ago

Idea: Run 3 AutoML loops in parallel: one that optimizes metric A, the other metric B, and the third metric A*B, using the same shared hyperparameter repository.

stale[bot] commented 1 year ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs in the next 180 days. Thank you for your contributions.