Part 5/9:
Runs on Flask, with minimal code (~25 lines).
Loads the pre-trained model and creates 512-dimensional vectors.
Responds rapidly (around 1/20th of a second per request).
The plan is to develop multiple microservices, possibly up to 10 or more, for functionalities including:
Long-term memory management
Speech recognition and synthesis
Natural language understanding
Search and retrieval
This modular structure aims to keep Raven lightweight, flexible, and open for community contributions.
Roadmap and Development Milestones
Several key documents and plans guide Raven's development:
- Release One Roadmap: Lays out the initial features and goals needed to consider the project's first version complete.