-
Notifications
You must be signed in to change notification settings - Fork 21
Open
Description
Currently, inference.py only runs a single hardcoded math problem. It would be useful to have an interactive chat mode where users can type math questions on the spot and receive instant solutions without needing to restart the script.
Problem
Users must edit inference.py every time they want to test a new question
Model reloads for each run (slow on limited hardware)
No continuous conversation experience
Proposed Solution
Add a new chat.py script that:
Loads the model once at startup
Runs an interactive loop accepting user questions
Generates step-by-step solutions in real-time
Allows users to exit gracefully with quit/exit commands
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
No labels