vh
3fec668bf2
- scripts/erp-tune-gx10/pull-verify-jenerallee78.sh + base-pin-jenerallee78-shards.txt: revision-pinned root-shard pull, 32/32 sha256+size vs brokkr-smithy pins, index set-equal to stock, STOCK tokenizer set installed over the repo's (which bakes in a 256-token truncation); repo originals kept as *.repo - scripts/erp-tune-gx10/run-06-gx10.json + launch-run-06.sh: run-05 config with the base swapped, recipe-r6, survivors-r5 verbatim, stock template path - docs/runbooks/gx10-run-06.md: pull/verify record, free-check result (encode reproduces run 5 exactly), hf download --include gotcha, gate naming (erp-seat-base-ara / erp-tune-v6)
48 lines
4.6 KiB
JSON
48 lines
4.6 KiB
JSON
{
|
|
"output_dir": "/home/infra-ops/erp-tune/run-06",
|
|
"roots_dir": "/home/infra-ops/erp-tune/datasets/derived",
|
|
"base_model_path": "/home/infra-ops/models/gemma4-26b-a4b-it-ara-abliterated-jenerallee78-0631379a",
|
|
"base_model_revision": "jenerallee78/gemma-4-26B-A4B-it-ara-abliterated @ 0631379a3d859e0059bc8d9b21ab5b654dfc272c (ARA abliteration of stock google/gemma-4-26B-A4B-it; 32 bf16 root shards sha256-verified against brokkr-smithy base-pin-jenerallee78.json; index sha256 33c59654e658a30fa29cdc87ccd6a752bfa0bb3e32cd56f95ff1eb82075e593a). THE SINGLE VARIABLE vs run 5: base only. Run-5 recipe, survivors, mask, template, hyperparameters all UNCHANGED. Tokenizer set = STOCK (tokenizer.json cc8d3a0c / tokenizer_config.json 9f4fec4b / chat_template.jinja ae53464b) copied over the repo's, whose shipped tokenizer.json carries a baked-in max_length=256 truncation; repo originals kept beside as *.repo. Runs 3/3c/4/5 were settled 2026-09-08 as STOCK base (index 907826a6), so this is the line's first abliterated base.",
|
|
"recipe": "/home/infra-ops/erp-tune/recipe-r6/recipe-erp-seat-sft-r6.json",
|
|
"survivors": "/home/infra-ops/erp-tune/recipe-r5/survivors-r5.jsonl",
|
|
"chat_template_path": "/home/infra-ops/models/gemma4-26b-a4b-it-bf16/chat_template.jinja",
|
|
"impersonation_mask_path": "/home/infra-ops/erp-tune/recipe-r3/lossmask-r3.jsonl",
|
|
"lora_rank": 64,
|
|
"lora_alpha": 128,
|
|
"lora_dropout": 0.0,
|
|
"max_seq_len": 16384,
|
|
"epochs": 1,
|
|
"seed": 20260824,
|
|
"per_device_batch_size": 2,
|
|
"gradient_accumulation_steps": 8,
|
|
"learning_rate": 0.0002,
|
|
"warmup_ratio": 0.1,
|
|
"lr_scheduler_type": "cosine",
|
|
"weight_decay": 0.01,
|
|
"load_in_4bit": false,
|
|
"gradient_checkpointing": true,
|
|
"loss_chunk_tokens": 1024,
|
|
"training_eligibility_override": "operator-2026-09-08-rnd-run6",
|
|
"overridden_blockers": [
|
|
"contamination-scan-not-implemented",
|
|
"stage-2-csam-detector-inert"
|
|
],
|
|
"substitute_controls": [
|
|
"pre-training holdout, run-1 (8,404 samples, work/card/session split)",
|
|
"pre-training holdout, govreport/holdout-v1 (416 reports, sha256-ranked, never_trained_on)",
|
|
"pre-training holdout, qmsum/holdout-v1 (5 transcripts, sha256-ranked, never_trained_on)",
|
|
"stage-A lexical quarantine, RP (829 records held unread)",
|
|
"stage-A lexical quarantine, run-5 slot (133 records held unread, /mnt/smithy/datasets/quarantine/r47-run5-longdep-screen/)",
|
|
"SCROLLS-membership disclosure on both slot sources (avoidance, NOT a scan): govreport + qmsum are SCROLLS/ZeroSCROLLS members, in no hoard/default-benchmarks.yaml entry and used by no R47 instrument",
|
|
"kvasir is HELD, not re-cut: the 1,613 kvasir survivors are reused verbatim from survivors-r4.jsonl (which cut run-3's seed-20260824 prefix at 3,347,622 ctx). survivors-r5.jsonl = survivors-r4 minus airoboros plus the govreport + qmsum roots whole; sha256 a25169a6258cd4abb0cb494a176a921c0e98eb73d65c53d033b6ee18293a43ae.",
|
|
"window_count belt-and-suspenders (SFT-RECIPE-run5-SCOPE.md 7.1): every govreport + qmsum row renders <= 14,000 tokens (max 9,385 / 13,700) and the harness never packs across samples, so window_count MUST be 1 on every slot row; a chunked_into_2 or single_window_truncated on either new root in truncation-report.json is a BUILD DEFECT and the run is killed before training.",
|
|
"SINGLE VARIABLE vs run 5: the BASE. Stock google/gemma-4-26B-A4B-it OUT, jenerallee78 ARA abliteration @ 0631379a IN. Corpus (survivors-r5 rows verbatim), impersonation loss-mask (lossmask-r3), stock chat template ae53464b, lr 2e-04, max_seq_len 16384, rank 64, alpha 128, dropout 0.0, cosine, warmup 0.1, wd 0.01, batch 2 x accum 8, 1 epoch, seed 20260824 all UNCHANGED from run 5.",
|
|
"FREE CHECK (brokkr-smithy, 2026-09-08): the [encode] pass must reproduce run 5 EXACTLY -- 8,370 records, ctx 18,598,779 tok, loss 9,935,076 tok -- because corpus, tokenizer and template are identical; any difference means the wrong tokenizer/template loaded and the run is killed before [train].",
|
|
"HOST: pfi-gx10 (GB10, aarch64, sm_121, 121 GB unified). Base shards sha256-verified against the revision-pinned HF LFS oids after landing; harness eitri-smithy 0a6bd2e; corpus COPIED, box mounts no NFS. Grant: operator directive to infra-ops in-session 2026-09-08 (\"unload the gx10 and commence training on the gx10. window is open now.\").",
|
|
"SURVIVORS: recipe-r6 ships no survivor list of its own (targets byte-identical to r5), so survivors-r5.jsonl (sha256 a25169a6...) is reused verbatim; recipe-erp-seat-sft-r6.json sha256 6499555471181bd8ef7a273340f162e44fe54bf4f070198f050c8be33ee769a8 from brokkr-smithy 4dd7590."
|
|
],
|
|
"unfittable": "drop",
|
|
"holdout_dir": "/home/infra-ops/erp-tune/datasets/holdout",
|
|
"save_steps": 50
|
|
}
|