Lessons

1The experiment is larger than the model file35 min read

Write a manifest that identifies a result and its dependencies.

  • →Record enough experiment state to reproduce a comparison
Read lesson
2A checkpoint is a continuation contract35 min read

Explain which state is required to resume an optimizer trajectory.

  • →Resume training with optimizer and data state
  • →Track data order and preprocessing versions
Read lesson
3A seed is one control, not a reproducibility guarantee35 min read

Design separate tests for repeatability and seed robustness.

  • →Control randomness without claiming universal determinism
  • →Separate reproducibility from robustness across seeds
Read lesson
4Prove the optimized implementation matches a reference35 min read

Design a small reference test for a vectorized operation.

  • →Verify an implementation against a small reference
Read lesson

Skills in this course

  1. 01Record enough experiment state to reproduce a comparisonRecord enough experiment state to reproduce a comparison.
  2. 02Resume training with optimizer and data stateResume training with optimizer and data state.
  3. 03Control randomness without claiming universal determinismControl randomness without claiming universal determinism.
  4. 04Verify an implementation against a small referenceVerify an implementation against a small reference.
  5. 05Track data order and preprocessing versionsTrack data order and preprocessing versions.
  6. 06Separate reproducibility from robustness across seedsSeparate reproducibility from robustness across seeds.