Lightlines
By orchestra-research

model-merging

orchestra-research892 installs

Blend language models without training again.

What it does

Creates merged language models that combine fine-tuned capabilities without retraining.

When to use it

Use when combining model capabilities, blending domain expertise, or rapidly experimenting with model variants.

How to use it

Give it model names or paths; it configures and runs a merge that saves a merged model locally.

What you provide

  • model names or paths

Uses

Latest review

What you need · 5

The model-loading example runs through Python; no runtime version is specified.

Runs merges with the mergekit-yaml command, supplied by the instructed mergekit installation.


Access · 4

This skill

HuggingFace Hub

Write

Uploads models to HuggingFace Hub

GitHub

Read

Downloads mergekit from GitHub