Skip to content

test: document QJL regression in test_turboquant_improves_over_polarquant#62

Open
brosequist wants to merge 1 commit intoTheTom:mainfrom
brosequist:test/qjl-regression-assertion
Open

test: document QJL regression in test_turboquant_improves_over_polarquant#62
brosequist wants to merge 1 commit intoTheTom:mainfrom
brosequist:test/qjl-regression-assertion

Conversation

@brosequist
Copy link
Copy Markdown

Summary

  • test_turboquant_improves_over_polarquant had no assertion — it only printed values. This adds a regression assertion documenting the empirical finding that QJL increases inner-product distortion (TQ 2-bit avg ≈ 0.091 vs PQ 2-bit avg ≈ 0.041).
  • The assertion is written as a regression guard: if QJL is ever fixed to actually improve things, the test will fail with a clear message prompting re-evaluation.

Test plan

  • pytest tests/test_distortion.py::TestDistortionScaling::test_turboquant_improves_over_polarquant

🤖 Generated with Claude Code

…uant

The existing test ended with a print() and no assertion, silently allowing QJL
to be worse than PolarQuant. This updates the test to assert the known finding:
QJL (TurboQuant 2-bit) is actively worse than MSE-only PolarQuant at the same
bit budget. The assertion will alert if QJL is ever fixed and starts winning,
prompting re-evaluation of the production path. See turbo4-resurrection.md.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant