Skip to content

Train from frozen scientific releases with verified progress and checkpoints - #51

Closed
MrScripty wants to merge 1 commit into
feature/scientific-cpu-comparison-20261013from
feature/scientific-frozen-training-20261013
Closed

MrScripty wants to merge 1 commit into
feature/scientific-cpu-comparison-20261013from
feature/scientific-frozen-training-20261013

Conversation

@MrScripty

@MrScripty MrScripty commented Oct 13, 2026 •

Copy link
Copy Markdown
Owner

Integration disposition — 2026-10-14

Current reconciliation (2026-10-14): qualified generic content from this original stack is superseded and integrated in merged PR56 on main, without unqualified PR43 ancestry. This original draft/branch remains retained as source/evidence and for any producer-v2-specific dependent guards. It is no longer a pending generic-feature merge. Do not merge the deferred legacy stack wholesale; its original actual-data acceptance remains open.

Original source and qualification record follows:

Tuldok can now train the pinned Rheon modal model from six reviewed, immutable Data releases through an editable graph. Frozen-release inputs preserve exact custody, while Runs displays bounded live progress, verified result metrics and a checkpoint download. The optional adapter consumes all 96 native observations without changing the original fitting or validation algorithms.

Actual Run 11fc2be1aada490f9b67c11c190728f6 completed and drained in 64.748344 seconds. Independent NumPy replay verified all input observations and 32,767 fitted gains (maximum relative difference 2.003e-14). Both validation trajectories and the untrained-model physics rejection passed; six artifact downloads and their exact bytes survived restart. Desktop/narrow browser checks passed, including a final read-only inspection with no new Run. An earlier failed diagnostic-prefix check is retained and covered by regression tests.

Validation: 26 focused generic checks, 10 producer tests (real eight-frame path), 12 training adapter tests with pinned inputs, 61 helper/import regressions, JavaScript controller/syntax checks, independent source and visual review, plan contract and diff checks passed. Final regression passed: 838 Python tests in 1,155.485 seconds; all five optional skips covered by 32 qualified reader/training tests with zero skips; 56 JavaScript syntax/controller checks passed. All 196 Python files and the final 123 JavaScript files remained unchanged during their respective checks. Both exact-head hosted workflows passed at 28c666f115ca8fac7daf29096e91493ff7b0cc10: UI run 38460254846 and lifecycle run 38460254594. See docs/plans/generic-scientific-projects/reports/p5b-verification.md for evidence and limitations.

Stacked on PR #50. Draft: do not merge. Fresh CPU fitting is qualified locally; autonomous evaluation with durable full fields is the next milestone. No GPU, new holdout, general 3D training, end-user scientific review or backup/restore acceptance is claimed here.

@coderabbitai

coderabbitai Bot commented Oct 13, 2026

Copy link
Copy Markdown

Important

Draft PR not reviewed

Draft PRs are not automatically reviewed by default.

  • Trigger a manual review

To automatically review draft PRs, update your CodeRabbit configuration:

reviews:
  auto_review:
    drafts: true
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

Copy link
Copy Markdown
Owner Author

Closing this original stacked proposal as selectively incorporated or superseded by merged #56 (generic scientific core) and #57 (generic scoped MCP).

Frozen scientific input capture, fresh CPU trainer/progress/checkpoints and original P5b receipts incorporated through PR56; current Pipeline/Runs provides broader exact retained composition.

Independent cleanup review checked this exact head 28c666f115ca8fac7daf29096e91493ff7b0cc10 against main ab2ca5ba225ae6f8fa91ce7dd743ba1e48ea72a0 and the admitted selective-integration lineage. Deferred controls-v2 and complete original P6/G7 scope remain preserved in open #43 and #53; this closure does not claim those gates passed.

The original branch is retained. Its head is not ancestral to main, so this cleanup neither deletes its history nor merges its excluded ancestry.

@MrScripty MrScripty closed this Oct 16, 2026
@MrScripty
MrScripty deleted the feature/scientific-frozen-training-20261013 branch November 20, 2026 01:52
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