BAI1013 · PROJECT EXPLORER

Move closer.
See the possibilities.

Three independent ideas.
Four windows into the one you could build.

Compare the scope and starting points

Hover, focus or tap a project to explore its repository and assignment scope.

CURRENT REPOSITORIES · PROPOSED ASSIGNMENT WORK

Source checked 10 October 2026. Project engines run locally.

Your next idea,
in four windows.

Move to a project above.
See its repository, workflow, engine and remaining work.

OPTION B

Class Interpreter

Reusable Python pieces · new connection

Explore project
01 · Repository

The current classroom interface.

main · 86362a3 · checked 10 October 2026

Repository-provided Class Interpreter interface image.

preview.png from the current repository. Its displayed text, counters and statuses are not live website results.

02 · Workflow

Speech to bilingual lesson notes.

faster-whisper → Argos Translate → summaries

  • Microphone or shared audio
  • English transcription / Chinese translation
  • Extraction or local DeepSeek summary

The browser interface and Python service are present in the repository. The model routes still need runtime verification.

03 · Engine

Processing stays with the local service.

Python · 127.0.0.1:8765

index.html / style.css / app.js
server.py / setup_models.py

Speech and translation need installed models. DeepSeek uses optional local Ollama. These services are not connected to this page.

04 · Assignment

Connect selection to a checkable summary.

Proposed work · NLP + AI API

  • Passage selection → model route
  • Supporting-ID validation
  • Notebook and checked evaluation

The current extraction and AI summary paths are alternatives; a source-linked selected-passage pipeline remains proposed work.

Class Interpreter · current repository, workflow, local engine and proposed assignment work.

THE REPOSITORY IS THE START

Pick the work.
Then make it real.

Keep the existing code, proposed scope and
engine requirements clearly in view.

Choose and view a plan

Class Interpreter is the recommended starting point for code reuse, followed by Yinwei and Fake News Detector. This is a planning judgement, not measured performance or a completed group decision. Choose one domain for both assignments, and demonstrate at least two connected AI techniques in a reproducible Python Jupyter Notebook.