Skip to content
Back to Deployed Systems
Project Brief

neurOS-v1 Package: neuros-sourceweigher — Data-Source Weighting

A focused service that weights heterogeneous data sources for training mixtures, so a foundation model learns from the right blend of data rather than from raw volume.

Neural Foundation ModelsNeural Data Infrastructure
neuros-v1data-weightingtraining-mixturesfoundation-models

The Right Blend, Not the Biggest Pile

neuros-sourceweigher is a small, sharp service that decides how much each data source should count when assembling a training mixture — the lever that keeps a foundation model from being dominated by whatever dataset happens to be largest.

package map

neuros-sourceweigher — data-source weighting

A focused service that weights heterogeneous data sources for training mixtures, so a foundation model learns from the right blend rather than raw volume.

Comprises

01

Weigher

source-weighting logic

02

Service

API around the weigher

Role

01

Mixture control

balance datasets for pretraining

02

Quality over volume

down- and up-weight sources

source weightingtraining-mixture control

What It Comprises

  • Weigher — the source-weighting logic that scores and balances data sources.
  • Service — an API wrapper so the weigher can be called from training and cloud pipelines.

How It Differs

It is the smallest package by surface area but a distinct concern: neuros-cloud moves and stores data and neuros-neurofm trains on it, while neuros-sourceweigher governs the mixture — turning "more data" into "better-balanced data" for pretraining.

Part of the neurOS-v1 platform.