release: add versioned binary manifest (W5) - #141
Draft
localai-bot wants to merge 11 commits into
Draft
Conversation
Materialize CLAIM-ENG-RELEASE-BINARIES-W5 after merged PR #129. No W5 implementation or artifact evidence is included in this checkpoint. FOLLOWING_AGENTS_PROTOCOL Assisted-by: Codex:gpt-5 [Codex]
Implement W5 with a deterministic stdlib-only manifest generator, versioned schema, canonical CPU/CUDA fixtures, and fail-closed mutation coverage. Advance only the release program lifecycle to ACTIVE while preserving every artifact and runtime gate as pending. FOLLOWING_AGENTS_PROTOCOL Assisted-by: Codex:GPT-5 [Codex]
Make schema constants JSON-type-strict, bind CPU tier inventories exactly, and add isolated coverage for the seven fresh-review production-removal mutations. Move readable golden fixtures into the PR-size-exempt test-script fixture tree without changing their bytes. FOLLOWING_AGENTS_PROTOCOL Assisted-by: Codex:GPT-5 [Codex]
Add independent removal tests for Python JSON bool/int identity and the required external NVIDIA driver declaration. Refresh the W5 evidence counts while keeping release artifacts and runtime gates pending. FOLLOWING_AGENTS_PROTOCOL Assisted-by: Codex:GPT-5 [Codex]
Preserve the reviewed release manifest implementation while taking current main's keyed records as the reconciliation base and union-appending both evidence histories. FOLLOWING_AGENTS_PROTOCOL Assisted-by: Codex:gpt-5 [Codex]
Preserve the reviewed release manifest implementation while taking e484a63 keyed records wholesale and union-appending both evidence histories. FOLLOWING_AGENTS_PROTOCOL Assisted-by: Codex:gpt-5 [Codex]
Restore the STATUS introduction from current main and compact the W5 release marker so the keyed page remains inside its size ratchet. Keep the release contract checker and mutation fixture pinned to the exact marker, with the required BENCHMARKS checkpoint in the same atomic commit. FOLLOWING_AGENTS_PROTOCOL Assisted-by: Codex:gpt-5 [Codex]
Take origin/main's keyed records wholesale, reapply only the W5 release clauses, preserve the reviewed manifest implementation byte-for-byte, and union the append-only evidence while removing inherited literal merge markers. FOLLOWING_AGENTS_PROTOCOL Assisted-by: Codex:gpt-5 [Codex]
Ignore workstation-only row branches so local preflight and clean CI classify the same shared evidence. Reconcile MODEL-EMBED-llama-llama-for-causal-lm to PARTIAL after its one-of-eight membership landing, with the real oracle gate still pending. FOLLOWING_AGENTS_PROTOCOL Assisted-by: Codex:gpt-5 [Codex]
Collaborator
Author
|
CI record-gate repair at
No release artifact/runtime/performance claim changed; reviewed W5 implementation blobs remain untouched. |
Recompute every engine lifecycle area from its matrix section so stale per-area counts fail closed while preserving the independently verified totals. FOLLOWING_AGENTS_PROTOCOL Assisted-by: Codex:gpt-5 [Codex]
Take current main's keyed benchmark and status records wholesale, reapply only the W5 release clauses, preserve the reviewed W5/audit/rollup blobs byte-for-byte, and retain main's append-only benchmark evidence, five-way plan, and CUDA build fix exactly. FOLLOWING_AGENTS_PROTOCOL Assisted-by: Codex:gpt-5 [Codex]
mudler
added a commit
that referenced
this pull request
Aug 8, 2026
The at-a-glance row, the section heading and the closed-row table all called the benchmarked model Laguna-XS-2.1. The measured checkpoint is poolside/Laguna-S-2.1-NVFP4: 118B total / ~8B active MoE, 48 layers, 256 experts, ~67 GiB. The label came from the local checkpoint directory being named laguna-xs-nvfp4. Evidence that the two names are one benchmark: the same 37.55 -> 44.46 vs vLLM 43.10 pair appears in this file under "Laguna-XS NVFP4" and in the same document's row for "Laguna-S-2.1 MoE (LagunaForCausalLM, 118B/8B)", both dated 2026-08-04; and the NVFP4 arm spec pins the checkpoint at poolside/Laguna-S-2.1-NVFP4, ~67 GiB, with layers 1..47 MoE. The section now states the model geometry and says where the XS label came from, so it cannot drift back. The reproduce row keeps the real directory name with a note that it holds the S-2.1 checkpoint. Numbers, ratios and evidence anchors are unchanged; this is a naming correction only. FEATURES.md and README are untouched: they list "Laguna-S / Laguna-XS 2.1" as a model family, which is a separate question from which checkpoint was measured. No open issue or PR covers this (searched issues and PRs for laguna naming; open PRs are #127, #128, #140, #141, none related). FOLLOWING_AGENTS_PROTOCOL Assisted-by: Claude Code:claude-opus-5 [ClaudeCode]
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
2 participants
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.
Claim\n\nImplements W5 of
ENG-RELEASE-BINARIESfrom the accepted release matrix merged in #129. This draft PR is the activeCLAIM-ENG-RELEASE-BINARIES-W5claim.\n\nClaim checkpoint:29107d0b. No W5 implementation or release artifact evidence existed when this draft was opened.\n\n## Scope\n\n- versioned manifest schema and deterministic build-time generator/validator\n- independent four-state evidence, per-CPU-tier and per-CUDA-SM metadata\n- canonical fixtures, mutation tests, and release-contract registration\n- required lifecycle and public pending-state checkpoints\n\nExcluded: W1-W4, W6-W13, archive/install/package/publish targets, runtime artifacts, GPU work, downloads, and services.\n\n## Current evidence\n\nClaim-only checkpoint; implementation and all artifact/runtime/correctness/performance evidence remain pending.