Repository navigation
docs(skill): record the HiCache outcome and the reset-path gotcha - #1362
Merged
Merged
Conversation
Judge the case as implemented rather than "worth integrating": HiCache ships as calibrators/hicache.py (#1359), the landing point moved from "a basis switch on Taylor" to the _basis_term hook plus a separate calibrator, and the same-harness numbers now exist (FLUX 1.7x: 33.33 dB vs reuse 31.72 and TaylorSeer-o2 30.91, sigma=1 collapses to 24.99 which reproduces the paper's sigma ablation). Also record two things the next integration should not rediscover: - a new calibrator's to_kwargs() must tolerate a null calibrator_kwargs, because refresh_context() rebuilds through empty(); TaylorSeer/DMD/SVD crashed there until #1359; - RandPipeline feeds the same hidden_states to every step, so a predictor's forecast equals the anchor and "the output differs" cannot prove the wiring. landscape.md gains the measured table (4.5): hit positions are identical across rows, and a mask-pinned pair keeps the ordering, so the gap is the basis and not the schedule.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #1361
Docs-only follow-up to #1359 (HiCache). The cache-algorithm skill still judged HiCache as "worth integrating, do it as a ~40-line
basisswitch on TaylorSeer", while the shipped implementation landed differently and produced same-harness numbers. A skill that keeps a stale verdict is worse than no verdict, so this records the outcome.Changes (3 files, documentation only)
SKILL.md_basis_term) instead of copying the accumulation loopreferences/landscape.mdσ=1is the undamped physicist-Hermite basis rather than a Taylor degeneracy); new §4.5 with the measured tablereferences/extension-points.mdto_kwargs()must bedict(self.calibrator_kwargs or {})—refresh_context()is the onlyreset=Truecaller andempty()nulls the field, which crashed TaylorSeer/DMD/SVD until #1359; points attests/api/test_calibrator_reset.py. Also records theRandPipelinetest trapThe measured table the skill now carries (§4.5)
FLUX.1-dev, 1024², 28 steps, seed 0,
rdt 0.12, 12/28 steps cached with identical hit positions across rows, 1× RTX PRO 5000, torch 2.15.0:Three reusable readings recorded with it: (1) at ~1.7× HiCache is the best of the three, which narrows the earlier "the
Bn_residuallayer is saturated" conclusion to a per-operating-point statement; (2)σ=1collapsing to 24.99 dB reproduces the paper's σ ablation and is the cheapest falsifier for "the basis is wired"; (3) a mask-pinned pair keeps the ordering, so the gap is the basis, not the hit sequence.Two traps now written down
to_kwargs()and the reset path.cache_dit.refresh_context(transformer, **shorthand)is the onlyreset=Trueentry, it rebuilds throughCalibratorConfig.empty()(every non-constant field becomesNone), and the rebuilt context callsCalibrator(config) -> to_kwargs().self.calibrator_kwargs.copy()crashes there; feat(cache): add HiCache damped-Hermite calibrator #1359 fixed the four sites and addedtests/api/test_calibrator_reset.py.RandPipelinecannot prove wiring by output diff. It feeds the samehidden_statesto every step, so the feature trajectory is constant and any predictor's forecast equals the anchor. Tests must monkeypatch a counter onapproximate()and assert the concrete state type / wired σ instead.Scope
Pure documentation: no
src/,csrc/,tests/or CI change; no runtime, performance or accuracy impact.pre-commit run --files <3 files>passes (mixed-line-ending, end-of-file).