Skip to content

Advance learned-provider placement and runtime 0.2 - #4

Merged
epi13 merged 2 commits into
mainfrom
agent/provider-offload-runtime-02
Aug 9, 2026
Merged

Advance learned-provider placement and runtime 0.2#4
epi13 merged 2 commits into
mainfrom
agent/provider-offload-runtime-02

Conversation

@epi13

@epi13 epi13 commented Aug 9, 2026

Copy link
Copy Markdown
Owner

Summary

  • add backend-neutral CPU/full-CUDA/sequential CPU offload policy with reserve, cap, workspace, dtype probes, and bounded AUTO recovery
  • add process-local Rust host lifecycle, reusable snapshots, bounded diagnostic normalization, measurements, quarantine, and executable HMM baseline
  • add optional Torch evidence adapter and bounded investigator context/workspace contracts
  • update schemas, roadmap, runtime architecture, threat model, and tests

Verification

  • Python: 38 tests, compileall, Ruff
  • Rust: fmt, Clippy -D warnings, 20 unit tests/doc-tests
  • demo ledger verification and provider benchmark
  • Joern baseline/post Rust query completed; parser produced empty method/call overlays, and Python frontend is unavailable in installed Joern

@epi13
epi13 marked this pull request as ready for review August 9, 2026 00:59
@epi13
epi13 merged commit 7c98f8f into main Aug 9, 2026
8 checks 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