llnl / llnl/professor

Add Parameter Scaling Options

Open
#11 0 comments 0 reactions 1 assignee View on GitHub

@cssherman is already working on this.

Since Jul 27, 2026.

enhancement
Dominant language
Python
Stars
2
Forks
1
Avg merge
4d 21h
Merged PRs (30d)
1

Description

Professor does not make any assumptions about the scaling of input parameters and output fields, which can cause problems during model training if their distributions vary significantly. Add an option to normalize values and improve the model learning rate / stability.

To achieve this, add a new user-facing option to prof-trainer called "scaling", which has the following options:

- none (default): Do not apply scaling to the dataset
- auto: Inspect the dataset before training and estimate the required scaling parameters
- filename: Read in pre-computed scaling parameters from a yaml format file

To validate the new option, do the following:

- Create a test dataset that has poorly scaled input parameters
- Compare model performance for both cases

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.