Skip to content

missing mention of setting up an embeddings model as obsidian-AI-providers extension allows (and local-GPT requires) #1621

@michahell

Description

@michahell

Prerequisites

Describe the issue

Hello,

In the following guide: https://openresource.dev/articles/running-a-local-ai-inside-obsidian-with-ollama/
(which is excellent) there are a few things outdated, namely:

  • there is a main model choice now when adding AI providers
  • gemma4:latest at the time of writing this issue does not support embeddings itself! gemm4embedding (and others) do.
  • the obsidian-AI-providers extension mentions specifically in it's own documentation to choose the local-nomic-embed-text model for this.

Currently, when following the guide, you will get errors. See my issue thread about this,
here: pfrankov/obsidian-local-gpt#95 and here: pfrankov/obsidian-ai-providers#26

What operating system(s) are you seeing the problem on?

macOS

What browser(s) are you seeing the problem on?

N/A

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions