Skip to content

Publish learned skill capsules and scene-specific installation - #12

Merged
Tarzelf merged 1 commit into
mainfrom
codex/skill-capsule-acquisition-20261007
Oct 8, 2026
Merged

Tarzelf merged 1 commit into
mainfrom
codex/skill-capsule-acquisition-20261007

Conversation

@Tarzelf

@Tarzelf Tarzelf commented Oct 8, 2026

Copy link
Copy Markdown
Contributor

Question

Advance the physical grounding research from task metadata plus an authored adapter to an installable learned movement artifact.

Evidence

This release distills the teacher's joint-target computation into a 34,171-byte canonical NumPy RBF model. Task phases, waypoints, grasp/contact checks and recovery remain authored. A new frozen eight-scene evaluation produces 7/8 for both teacher and student, without teacher fallback; total episode wall time increases from 6.273 to 6.360 seconds. Historical benchmark inputs, physics, success rules and published evidence remain unchanged.

A 39,845-byte data-only capsule binds task source, grounding, model and ten runtime source files. The local studio imports it, runs native placement and identical arm-command replay with open jaws, and enables execution only for that exact qualified scene. Browser verification includes file import, scene-change invalidation, requalification and execution without retraining. The bundled button preserves exact JSON bytes; rejected replacement imports clear readiness.

One known default scene plus repeat passed native macOS network denial after a direct libc connection was rejected with EPERM. Extracting the final archive reproduces exact actions, sampled states, metrics and native success checks with preinstalled dependencies. These are known-scene integration checks, not unseen/general robot competence or calibrated physics.

Reproduction

The new report and source/evidence live under topics/physical-grounding/runs/2026-10-07-skill-capsule-v1/. Download/extract release/simlab-v0.4.zip, install requirements.lock.txt once, then run python -m simlab.studio --offline and drop the bundled capsule. CAPSULES.md covers author/train/build/install/qualify/run. native-offline/check.py reproduces the native deny-network check on macOS.

The release includes the dataset, protocol, collector/experiment source snapshots, frozen evaluation, profiling, native controls, browser replay, manifest and final-archive reproduction receipt. No GPU, hardware execution, learned perception or human-video motor supervision is claimed.

Checks

  • Primary sources support the methodological claims; local receipts supply measurements.
  • No private footage, credentials, personal data or copied paper figures are included.
  • Original code, generated teacher data and movement coefficients use MIT; dependencies retain their licenses and are not bundled.
  • Product suite: 124 passing tests; affected studio/installer checks pass after UI changes.
  • Native network denial and extracted-source exact reproduction pass.
  • Research build, validation and publication integrity tests pass.

@Tarzelf
Tarzelf merged commit 632be79 into main Oct 8, 2026
1 check passed
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