FreeCampus Python

Milestone Challenge: Write a Model Release Decision

Use a reproducible experiment to justify a narrowly scoped classroom-demo decision.
Open in Colab
  • Goal: Use a reproducible experiment to justify a narrowly scoped classroom-demo decision.
  • Estimated time: 2–4 hours of active work.
  • Before starting: Complete the milestone units and their unit challenges.
  • Keep: Your source, observed outputs, checks, and a debugging note.

This community challenge is ungraded. Use the checks to review your own work; there is no submission service or automatic grade. Completion is self-reported in this browser on this device.

1. Write a model release decision

Reuse the Iris baseline and shallow-tree experiment. The proposed product is a classroom classification demo, not a scientific identification service. Write a model card with task, data, split, metrics, errors, intended use, prohibited claims, and limitations. Compare both models and recommend release, revise, or stop for that narrow demo.

Predict an ordinary result before coding. Start with the smallest part you can run and inspect. Use Colab for notebooks and the local workspace introduced in the units for packaging or command-line tasks.

2. Set up the experiment

decision_card = {
    "intended_use": "classroom demonstration",
    "data_source": "scikit-learn Iris dataset",
    "split": {},
    "baseline_results": {},
    "model_results": {},
    "limitations": [],
    "recommendation": "",
}
# Populate the card from reproducible outputs and analysis.

The starter is deliberately incomplete. Implement the contract; do not change the expected results to match an incorrect implementation.

3. Check the observable results

  1. A fresh run reproduces recorded partitions, parameters, and metrics in the declared environment.
  2. Include sample counts and confusion matrices, not accuracy alone.
  3. Support every claimed advantage with a recorded measure or explicitly scoped observation.
  4. Describe an input outside training conditions and why reliable identification cannot be promised.
  5. State whether this dataset contains personal data and why collecting photographs or learner data would require a new privacy review.

Record actual outputs beside predictions. Reduce a failure to a small example, change one cause at a time, and repeat the failed check after repairing it.

4. Change one condition

A volunteer proposes a transformer and a general plant-identification service. Identify missing data, evaluation, consent or licensing questions, and operating limits. Explain why changing architecture cannot supply missing evidence.

5. Review a hint after attempting the task

Hint: connect the relevant tools

Treat the unit notebook as the evidence source. Separate observations from future work, compare conditions, and name who is affected by errors.

6. Explain and self-review

Deliver the notebook, model card, and one-page decision. Ask a learner to reproduce one measure and challenge one claim, then revise unsupported claims. This exercise does not certify a deployed model.

Review path: compare your result with the contract

Check the ordinary result first, then the boundary cases and changed condition. Explain which rule justifies each result. An implementation that differs from the hint can still satisfy the contract; explain why yours does.

7. Save your work and record completion

Save the notebook or project somewhere you control. Include one decision, one failed check, its repair, and a result you can reproduce. Record completion after reviewing the checks; you can undo it later.

Not yet recorded.

Return to the milestone overview.

Back to top