Skip to content

More robust Docker container #35

@deadbits

Description

@deadbits
  • Vector database should be a persisted volume when using Docker so you don't have to download and load the data every time. Also, any auto_update added records will be lost between runs right now.
  • Modify entrypoint process so other embedding models can be used. right now its just openAI because they are hard-coded in entrypoint.sh
  • Use Docker compose

Metadata

Metadata

Assignees

Labels

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions