clarke / submission_checklist.md
yashvshetty's picture
Clarke: NHS clinical documentation system
9636a02

Clarke — Submission Checklist

Ready Now (End of 24-Hour Build)

  • Public HF Space: [URL pending] — accessible, runs 3 demo scenarios
  • Public GitHub repo: [URL pending] — clean code, README, licence
  • Public LoRA adapter: [Not published] (if trained) — traces to google/medgemma-27b-text-it
  • evaluation_report.md — metrics computed (if completed)

Buffer Week (Mon 16 – Sun 22 Feb)

  • 3-page writeup (Mon 16 Feb)
  • 3-minute video (Tue 17 – Sat 21 Feb)
  • Final Kaggle submission (Sun 22 Feb)
    • Writeup submitted via Kaggle Writeups tab
    • Agentic Workflow Prize selected
    • All links included: video, GitHub, HF Space, HF model

Verification Notes (Task 37 run)

  • HF Space and GitHub public URLs are not present in this repository snapshot; external incognito validation could not be completed in this environment.
  • evaluation_report.md exists in repository root.

Task 38 — Phase 5 Final Checkpoint (Current Repository Snapshot)

Day 3 Criteria Verification

  • Public HF Space is live and accessible from any browser. Status: FAIL (no public HF Space URL is present in README.md or this checklist).
  • Public GitHub repo with clean code, README with architecture diagram, and docstrings. Status: PARTIAL/FAIL (README architecture diagram and licence are present, docstring scan passes, but no public GitHub URL is provided).
  • At least 3 demo scenarios work flawlessly on the live HF Space. Status: FAIL (no live HF Space URL available for external verification).
  • Submission checklist confirms all competition requirements are met (except video and writeup). Status: FAIL (Ready Now section still has unchecked public artefact items).

Nice-to-Haves Inventory (for writeup/video planning)

  • LoRA adapter trained and published? N (adapter publication skipped in current repo state).
  • WER comparison table? Y (WER metrics recorded in evaluation_report.md).
  • EHR Agent metrics? Y (fact recall/precision/hallucination metrics recorded).
  • BLEU/ROUGE-L evaluation? Y (BLEU and ROUGE-L recorded).
  • evaluation_report.md populated? Y.
  • UI visually polished? Y (recorded as completed in Task 33 section of evaluation_report.md).

Commands Used

rg -n "URL pending|add your public Space URL|<your-github-repo-url>" README.md submission_checklist.md
grep -rL '"""' backend/ frontend/ --include='*.py'