DS 6210 — Spectral Optimizer Capstone

Course materials and starter code for the DS 6210 take-home capstone on spectral optimization for deep learning.

Layout

project/
  spectral_harness/           starter code -- read its README first
    README.md                 quickstart, rules, self-audit checklist
    my_optimizer.py           student work lives here
    harness/                  fixed scaffolding (model, data, timing, seeds)
    experiments/              4 scripts: baseline, your method, parity, grid
    tests/                    pytest parity contract
    reproduce.sh              end-to-end smoke

Quickstart

cd project/spectral_harness
pip install -r requirements.txt
./reproduce.sh

The full assignment sheet (problem statement, rubric, rules) is distributed separately by the instructor. Read project/spectral_harness/README.md for the starter API and the self-audit checklist.