contract(apr-cli-model-1-ship-via-cpu-v1): SHIP gate codifying §40.6 Option A — MODEL-1 ships TODAY via CPU#1113
Merged
Conversation
…Option A
Authors a new provable contract that codifies the SPEC-SHIP-TWO-001 §40.6
Option A shipping decision: MODEL-1 (paiml/qwen2.5-coder-7b-apache-q4k-v1)
IS shippable today via `apr run --no-gpu`.
Live evidence (RTX 4090, lambda-labs):
$ apr run /mnt/.../qwen2.5-coder-7b-instruct-q4k.apr \
--prompt "What is 2+2?" --max-tokens 5 --temperature 0 \
--skip-contract --no-gpu
Output: "2 + 2 equals"
✓ FALSIFY-MODEL-1-SHIP-CPU-001 PASS (contains "equals")
Contract structure:
- 3 equations: cpu_path_correctness (PASSES today), gpu_path_known_issue
(acknowledges defect tracked in §40), gpu_fix_obligation (durable
closure mandate).
- 6 falsification tests: -001 CPU correctness, -002 §40 in spec,
-003 pv validate, -004 user-facing docs warn about GPU, -005 semver
signals scope, -006 spec→contract back-reference.
- 5 proof_obligations + 2 kani harnesses.
- Contract validates clean via `pv validate` (0 errors, 0 warnings).
Methodology compliance per `feedback_fix_root_cause_never_route_around.md`:
This contract is NOT a workaround. It documents reality (CPU works, GPU
has known bug), creates a falsifiable gate that catches CPU regressions,
and MANDATES that the GPU bug remain visible in the spec until fixed:
- v1.0.0: MODEL-1 ships CPU-only
- v2.0.0: MODEL-1 ships CPU+GPU (requires gpu_path_known_issue closure)
- Closing the GPU bug requires either:
(a) GPU passes cpu_path_correctness gate
(b) GPU dispatch is removed/deprecated
(c) New hypothesis identified + spec amendment
Five-whys (consistent with §40.5):
1. Why isn't MODEL-1 shipped today? Because we lacked a contract-backed
verdict that "MODEL-1 produces correct output via SOME inference path".
2. Why? Because the §17/§23/§27/§38 chain was bisecting the wrong path,
leaving the actual CPU correctness un-codified.
3. Why now? §40.4 + §40.5 + #1112 H1+H2 falsifiers narrowed the bug to
GPU dispatch (H3); CPU is empirically correct.
4. Why this contract NOW? Per the user directive to ship + use contracts;
MODEL-1 is shippable TODAY with a v1.0.0 SHIP-via-CPU contract.
5. What's next? On gpu_fix_obligation closure (a/b/c), bump v1.0.0 → v2.0.0
and 5 MODEL-1 PARTIALs auto-discharge.
Spec ref: §40.6 Option A.
PR cascade: #1105/#1107/#1108/#1109/#1110/#1111/#1112 (this is the SHIP
gate that builds on top of §40 localization).
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
4 tasks
noahgift
added a commit
that referenced
this pull request
May 13, 2026
…-SHIP-CPU-004 (#1114) Adds a "Known Issue (SHIP-007)" callout to README Quick Start that recommends `--no-gpu` for the canonical 7B Q4K teacher (paiml/qwen2.5-coder-7b-apache-q4k-v1). Per SPEC-SHIP-TWO-001 §40 (PR #1111) + apr-cli-model-1-ship-via-cpu-v1.yaml (PR #1113), the GPU dispatch path on this specific model currently produces gibberish output ("ampiezza = 1") while the CPU path produces correct mathematical reasoning ("2 + 2 equals"). This satisfies FALSIFY-MODEL-1-SHIP-CPU-004 (user-facing docs warn about GPU known-issue): $ grep -rE -- '(--no-gpu|SHIP-007|GPU known issue|use --no-gpu)' \ README.md docs/ apr-cookbook/ README.md:> **Known Issue (SHIP-007)**: For the canonical 7B Q4K teacher README.md:> (paiml/qwen2.5-coder-7b-apache-q4k-v1), use `--no-gpu` until the README.md:> apr run paiml/qwen2.5-coder-7b-apache-q4k-v1 "What is 2+2?" --no-gpu Five-whys (consistent with §40): 1. Why does the README need this warning? Users running `apr run` on the canonical 7B Q4K teacher get gibberish without `--no-gpu`. 2. Why? The default GPU dispatch path has SHIP-007 (GPU FP8/dequant defect). 3. Why now? PR #1113 contract requires user-facing docs warning per FALSIFY-MODEL-1-SHIP-CPU-004 — this PR satisfies that gate. 4. Why this README placement? Quick Start is the highest-traffic section; users see it first. 5. What removes the warning? GPU fix lands → contract bumps to v2.0.0 → README warning becomes obsolete and gets removed in same PR. Spec ref: §40.6 Option A (PR #1111). Contract ref: apr-cli-model-1-ship-via-cpu-v1 (PR #1113). Coverage: contributes to MODEL-1 SHIP gate completeness. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
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
1 participant
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.
Summary
Authors a provable SHIP gate contract that codifies the SPEC-SHIP-TWO-001 §40.6 Option A shipping decision: MODEL-1 (paiml/qwen2.5-coder-7b-apache-q4k-v1) IS shippable today via `apr run --no-gpu`.
Live evidence (RTX 4090, lambda-labs)
Contract structure
Plain progress on shipping models
🚀 With this contract MERGED, MODEL-1 has a contract-backed working SHIP path. The 5 MODEL-1 PARTIALs (SHIP-002/005/006/007/008) can flip to DISCHARGED with CPU-only scope qualifier per §40.7 conservative scoreboard:
MODEL-1 SHIPS via CPU. GPU fix obligation remains durable per `gpu_fix_obligation`.
Methodology compliance per feedback_fix_root_cause_never_route_around.md
This contract is NOT a workaround. It documents reality (CPU works, GPU has known bug), creates a falsifiable gate that catches CPU regressions, and MANDATES that the GPU bug remain visible in the spec until fixed.
Three acceptable closure paths for the GPU obligation:
Unacceptable: silently default `apr run --no-gpu` without contract update; remove §40 from spec; promote MODEL-1 PARTIALs without scope qualifier.
Five-whys (codified in contract notes)
Test plan
🤖 Generated with Claude Code