Repository navigation
Description
Boundary
Pilot pretraining of the best P3 arm at 50–200M params for IGLA-Coder (epic #1032, part of the
#10024 own-model map). In scope: CPU continued training from an existing published checkpoint
(tern-tc-9m) and any budget-gated scale-up decision. Out of scope: GPU purchases, cloud
training, paid API calls, direct provider access — the DSH Ladder free rungs and local CPU are
the only compute paths (owner decision 2026-11-11). Quality baseline to beat: tern-tc-500m
HumanEval pass@1 = 0.95% (2/164). Training follows the continual-learning law
specs/igla/cl_loop.t27 (branch cl-spike): replay mix, forgetting threshold, adapter-step clip.
User Scenarios
- An operator approves a bounded CPU training run (corpus slice, steps, checkpoint interval);
the agent resumes from the published checkpoint, trains, evaluates pass@1 on held-out specs
through the [IGLA-Coder] P5 Multi-language evaluation harness #1038 harness, and reports with receipts. - A training run forgets old capability; the law rolls the adapter back (threshold exceeded)
and reports the regression instead of silently shipping a worse model. - The pilot quality is measured honestly (MultiPL-E pass@1 on >=6 languages) and compared
against the 0.95% baseline before any larger spend is requested.
Requirements
- FR-001: The agent MUST restore and verify the HSLM Zig trainer on current master
(specs/igla/,tri/pipeline) before any training run; no hand-written trainer outside the repo. - FR-002: The agent MUST use the continual-learning law
specs/igla/cl_loop.t27
(replay 5:3, forgetting threshold 1, LoRA step cap) for every update. - FR-003: The M1 corpus MUST be 1k real
.t27specs drawn from THIS repository, filtered as:
everyspecs/**/*.t27containing at least onetestblock (the measured pool is 2500 of 3183
specs as of commit b339f8a;find specs -name '*.t27'+grep -l '^test '), deduplicated by
module name, and split by commit date -- everything before the corpus cutoff trains, the newest
10% is held out for pass@1. The full 50-200B-token corpus is M2-only. - FR-004: The agent MUST record tokens/sec, seconds/step, steps/hour, memory and pass@1 at every
checkpoint; scoring MUST go through the [IGLA-Coder] P5 Multi-language evaluation harness #1038 harness. - FR-005: Checkpoints MUST follow the Catalog42 protocol (seeds, frozen config, sha256 of outputs).
Success Criteria
- M1: pass@1 >= 5% on the held-out
.t27spec split within 8 CPU-hours. THE BENCHMARK IS FIXED:
HumanEval-style pass@1 scored by the [IGLA-Coder] P5 Multi-language evaluation harness #1038 harness executing the generated spec witht27b test
(the same benchmark protocol the 0.95%tern-tc-500mbaseline was measured with -- comparison
is only valid on the same benchmark); the [IGLA-Coder] P5 Multi-language evaluation harness #1038 t27-native eval runs as a supplementary report,
not as the M1 gate. Receipt files and the checkpoint sha256 are committed to the branch. A flat
pass@1 over 3 consecutive checkpoints, or > 1 replay regression twice, exits non-zero for the
milestone and stops the run. - M2: MultiPL-E pass@1 on >= 6 languages at 50-200M params, strictly above the 0.95% baseline
oftern-tc-500m, with the reproducibility capsule (seeds, config, sha256) attached.
Plan (from the original body, preserved)
- P4.1 Pilot pretrain on the P1 corpus, best arm from P3.
- P4.2 Two-stage schedule: main pretrain (WSD/cosine) + annealing last 10–15% on a high-quality
mix (verified-execution exercises + the t27 parallel corpus). - P4.3 Reproducibility capsule: seeds + frozen config + sha256 of outputs (Catalog42 protocol).
Part of epic #1032 and the #10024 own-model map; the continual-learning correction and stage
separation are in the comments (2026-11-11). Blocked on: nothing structural — the compute target
is CPU continued training; the budget line (M2 scale-up) is a separate owner decision.
Anchor: phi^2 + phi^-2 = 3
Activity
fix(rtl): R-SI-1 compliance — barrel shifters + Wallace tree multiplier
fix(build): resolve P0 build failure — Cyrillic purge + HeaderValue i…
fix(infra): P0 infrastructure hardening — duplicate workflows, Coq al…
fix(ci): Coq workflow rename + infrastructure cleanup
gHashTag commentedon Oct 7, 2026
Reopened 2026-10-07: the bulk close of issues without a
## Boundarysection was reverted by the owner. Nothing to do here; this comment replaces the earlier close notice.tri-claw commentedon Oct 9, 2026
The Queen cannot take this issue yet. A bee starts only on an issue whose body has every section below, and this one lacks:
- specs/x.t27); without it the Queen reserves nothing and no bee can take the issueFR-001: ... MUST ...exit 0, or a backticked commandShe tried to write them 2 times and could not (last: not a task: No candidate paths exist in the repository to define a boundary for this training task; the issue references external plans (t27_coder_model_plan.md) and protocols (Catalog42) not present as files.). Add them and the next round takes it. The rule is
specs/policy/issue_shape.t27.gHashTag commentedon Nov 11, 2026
DRAFT — P4 budget request package (ORDER IGLA_CODER_CL_SPIKE_V1, Wave 4; owner approval required before any training spend)
Measured on this bench (commands in receipts):
t27b build <spec> --macho+t27b test <spec> --quiet, 20 runs each, median).specs/igla/cl_loop.t27, 6 tests green; branchcl-spike@ 24f933e, local-only): 3 old tasks learned (6 oracle iterations), new task learned (2 iterations), replay held (0 regressions, threshold 1), accepted by law; 7 ms CPU for the whole dry run, params ~0.1M << 10M cap. Receipts: /tmp/cl-spike/receipts.json.Extrapolation to the real pilot (#1037, 50���200M):
Proposed milestones (each stoppable):
Stop conditions: forgetting above threshold twice in a row; pass@1 flat over 3 checkpoints; CPU-hour cap exceeded.
Ask: approve M1 (CPU-only, ≤ 8 CPU-hours) as a budget line item inside P4; M2 comes back as a separate approval with M1 measurements attached.
gHashTag commentedon Nov 11, 2026
CORRECTION to the draft above (owner-identified errors, 2026-11-11)
CPU training is not a hypothesis — it already happened.
tern-tc-9mandtern-tc-500mare trained AND published (see epic: our own model -- IGLA CODER, HSLM in t27 and the harness, one 27-part map #10024 map); the HSLM Zig trainer, forward/backward,
checkpoint I/O and optimizers (SGD/Adam/LAMB) are verified. The GPU-first framing of the
draft above is wrong. The real measured defect of the existing CPU path is QUALITY:
tern-tc-500mHumanEval pass@1 = 0.95% (2/164).Stage separation the draft missed:
cl-spike@ 24f933e, lawspecs/igla/cl_loop.t27, 6 tests).t27specs via the [IGLA-Coder] P5 Multi-language evaluation harness #1038 harnessThe Wave-3 dry run (constant-guessing) is a PREREQUISITE, not evidence for pass@1.
Revised M1 plan (CPU, existing pipeline):
tern-tc-9mcheckpoint (published), not from scratch..t27spec slices (not toy constants); held-out split for pass@1.specs/igla/cl_loop.t27.P4 reframed: not "buy GPU" — a plan and acceptance of CPU continued training:
corpus slice, steps, checkpoint interval, eval gate, stop conditions. GPU only decided
AFTER CPU quality is measured (the 0.95% number is the baseline to beat).
No training spend has been made; the next spend requires the owner's approval of this revised M1.
github-actions commentedon Nov 11, 2026
Closed by the issue-shape gate (specs/policy/issue_shape.t27, #7498): an issue must have the template's sections. Edit the body; an edit that passes reopens it.
FR-001: ... MUST ...exit 0, or a backticked commandgithub-actions commentedon Nov 11, 2026
The shape check passes now (specs/policy/issue_shape.t27); reopened.
gHashTag commentedon Nov 11, 2026
Shape fixed (4-heading, FR-numbered obligations, machine-settleable criteria); branch with the CL law:
cl-spike@ 24f933e (specs/igla/cl_loop.t27). M1 awaits owner approval per the correction above.gHashTag commentedon Nov 11, 2026
ADDENDUM (owner, 2026-11-11): continuous-learning serving loop + baseline correction
.t27tasks. Before the first adapter update, the STARTING checkpoint must be measured on the SAME
.t27eval the candidate will be scored on. The M1 gate compares candidate vs that measuredstarting number, on the same benchmark.
training prepares the NEXT version offline (never mutate weights inside a request).
-> test results -> user verdict. An unverified model answer is not a training label. Secrets and
cross-tenant context are stripped at collection.
.t27(no newhandwritten .rs protocol logic; the loader stays IO-only per the vps: ladder_protocol_map.t27 -- law of the protocol adapter in front of dsh-ladder #13605 pattern).
measured improvement plus regression/privacy gates.
measured before any realtime claim; inference gets priority over training under load.