Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Store all local models in /models folder, with overridable model path #242

Closed
lalalune opened this issue Nov 9, 2024 · 1 comment
Closed
Labels
enhancement New feature or request

Comments

@lalalune
Copy link
Member

lalalune commented Nov 9, 2024

We should make sure that all models stored by llama, whisper, embedding, etc all get stored in a /models folder that is easy to find and clean. This way the repo doesn't get huge and locating models isn't messy.

Here are the steps:

  1. Look for all services where models are being loaded, including llama, fastembed, whisper, etc
  2. Change all paths to be eliza/models
  3. Add a path checker if that doesn't exist with some sane defaults to handle any case
  4. Add a model path override to functions so the packaged version doesn't go somewhere weird.
@lalalune lalalune added the enhancement New feature or request label Nov 9, 2024
@dr-fusion
Copy link
Contributor

Shouldn't we just make it an environment variable so users might have their models stored at some different directory for any other purpose and they can easily point to that directory? and or Downloading model happens to that particular path?

For example like the follow one.

in .env

#LlamaLocal Configuration LLAMALOCAL_PATH= # Default: "" which is the current directory in plugin-node/dist/ which gets destroyed and recreated on every build

then in llama.ts line no 188,
this.modelPath = path.join(process.env.LLAMALOCAL_PATH, modelName);

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement New feature or request
Projects
None yet
Development

No branches or pull requests

2 participants