SPEAR Logo
0.3.0
  • 1. Getting started
    • 1.1. Getting started
    • 1.2. Running the images
    • 1.3. Installation
    • 1.4. Container
  • 2. Overview
  • 3. Using SPEAR
  • 4. Request modes and evidence
  • 5. The confined execution path
  • 6. The model and its weights
  • 7. Operations and reference
  • 8. Contributing
SPEAR
  • 1. Getting started

1. Getting started

Two ways to run SPEAR: a container that carries everything, and a native installation. Start with the quick start. A colleague who only wants to use SPEAR needs nothing past Running the images; the later chapters are for whoever installs it natively or builds the images.

  • 1.1. Getting started
    • 1.1.1. Prerequisites
    • 1.1.2. Install
    • 1.1.3. Configure a model backend
    • 1.1.4. Use a project
    • 1.1.5. Make a change
    • 1.1.6. Bind a standard
    • 1.1.7. Make a change the standard governs
    • 1.1.8. Read the result
    • 1.1.9. Retrieval is not a detail
    • 1.1.10. Where next
  • 1.2. Running the images
    • 1.2.1. What you need
    • 1.2.2. Getting the image
    • 1.2.3. Running the reds image
    • 1.2.4. Laying out your checkouts
    • 1.2.5. Running it
    • 1.2.6. The first session: indexing
    • 1.2.7. When something is wrong
  • 1.3. Installation
    • 1.3.1. Prerequisites
    • 1.3.2. Native installation
    • 1.3.3. Container installation
    • 1.3.4. The inference runtime
    • 1.3.5. What ends up where
    • 1.3.6. Checking the installation
  • 1.4. Container
    • 1.4.1. Managing images
    • 1.4.2. Daily use
    • 1.4.3. What is baked, and what is not
    • 1.4.4. Two build contexts
    • 1.4.5. The optional inputs
    • 1.4.6. Three profiles
    • 1.4.7. Baking the trees
    • 1.4.8. Common and local state
    • 1.4.9. A deployment’s own content, and how it gets in
    • 1.4.10. Publishing, and not publishing
    • 1.4.11. Relative corpus paths
    • 1.4.12. Mounts are derived, not listed
    • 1.4.13. The three flags that are not optional
    • 1.4.14. Refreshing the index
    • 1.4.15. Design rationale: retrieval stays in the client
Previous Next

© Copyright 2026, HEIG-VD/REDS. Last updated on 21 Nov 2026, 16:10.