conf30: builds and packs the training stream, builds the FVT transplant of Qwen2.5-1.5B, runs the gates, trains 1B tokens with checkpoint scoring, and scores the final checkpoint per document.

Succeeded · Registered by @stw2 via agent. Reported and received times are kept apart; no computation or result is verified.

Pinned source and configuration

https://github.com/stw2/tokenizer-science-tax @ 16a182f7c3592e9e2e84c5a9bc824d2e9b263a06

Reference checked 2026-09-14 13:12 UTC. Later commits, branches or plan changes do not retarget this attempt.

experiments/E06-math-code-rescue/scripts/00_bootstrap_cuda.sh

Command: bash scripts/00_bootstrap_cuda.sh conf30 configs/e6_conf15.json
Working directory: experiments/E06-math-code-rescue
Configuration paths: None
Parameters: config configs/e6_conf15.json; arm conf30: shares of APT4 tokens Polish 0.56, English 0.14, OpenWebMath 0.15, Python code 0.15; base model Qwen2.5-1.5B; 1B tokens, sequence 2048, micro-batch 8 x gradient accumulation 32, learning rate 1e-4, 50 warmup steps, no decay, weight decay 0.1, clip 1.0, checkpoints every 100M tokens; stream seed 42, shuffle buffer 10,000, documents of at least 200 characters, pool headroom 1.03; scoring: first 200 documents per text, digit probe as frozen in configs/probe_digits_spec.json; original data-manifest directory 44292676 renamed to conf30
Environment: 3.10 (conda-based PyTorch container image); PyTorch CUDA, bf16, sdpa attention, gradient checkpointing, bitsandbytes 8-bit AdamW, Liger fused cross-entropy; HF_HUB_DISABLE_XET=1; 4090-class GPU, 24 GB
Output directory: Not recorded

Inputs

Materials the registrant named when registering this attempt, by content identity; obtainability is derived from their location reports. Nothing is fetched or verified.

Qwen2.5-1.5B · base model · M124 Qwen2.5-1.5B · checkpoint · Download
APT4 tokenizer · tokenizer · M46 Bielik-PL-11B-v3.0-Instruct · model · Ask the reporter
packed stream, 30% math and code · training data · M192 Packed recovery pretraining stream, 30% math and code · training data · Ask the reporter
Polish FineWeb2-HQ evaluation holdout · evaluation data · M129 pl_eval.jsonl · raw output · Download
English SlimPajama-6B evaluation holdout · evaluation data · M130 en_eval.jsonl · raw output · Ask the reporter
math_clean synthetic arithmetic set · evaluation data · M128 math_clean evaluation statements · dataset · Download
en-arxiv-abstracts.jsonl · evaluation data · M27 en-arxiv-abstracts.jsonl · raw output · Download
en-latex-methods.jsonl · evaluation data · M28 en-latex-methods.jsonl · raw output · Ask the reporter
en-python-code.jsonl · evaluation data · M29 en-python-code.jsonl · raw output · Ask the reporter
en-gsm8k.jsonl · evaluation data · M30 en-gsm8k.jsonl · raw output · Download
pl-science-pes.jsonl · evaluation data · M31 pl-science-pes.jsonl · raw output · Ask the reporter
pl-wiki-science.jsonl · evaluation data · M32 pl-wiki-science.jsonl · raw output · Download
pl-informal.jsonl · evaluation data · M33 pl-informal.jsonl · raw output · Ask the reporter
synthetic arithmetic probe items · probe items · M63 probes.jsonl · raw output · Download

Delivered events

Registered
#1
conf30: builds and packs the training stream, builds the FVT transplant of Qwen2.5-1.5B, runs the gates, trains 1B tokens with checkpoint scoring, and scores the final checkpoint per document.
reported 2026-07-10 09:03 UTC · received 2026-09-14 13:13 UTC · @stw2 via agent · posted to the Thread
Started
#2
Started.
reported 2026-07-10 09:03 UTC · received 2026-09-14 13:13 UTC · @stw2 via agent · attempt only
Succeeded
#3
Completed with exit code 0.
Exit code 0.
- data_manifests/conf30/raw_manifest.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/data_manifests/conf30/raw_manifest.json · public · sha256 f12ef66fdc3b… · 1687 bytes · M314
- data_manifests/conf30/conf30.meta.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/data_manifests/conf30/conf30.meta.json · public · sha256 1fee85f605d3… · 604 bytes · M315
- FVT transplant model.safetensors · Not kept: rebuilt deterministically by the init stage of scripts/00_bootstrap_cuda.sh · unavailable · sha256 ebdaa65f2d8e… · M287
- env_manifest_conf30.txt · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/env_manifest_conf30.txt · public · sha256 c97323a10d45… · 6611 bytes · M288
- bpb_conf30.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/bpb_conf30.jsonl · public · sha256 f8f946d39ef6… · 14243 bytes · M316
- probe_digits_conf30.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_conf30.jsonl · public · sha256 5ebd75a205eb… · 11064 bytes · M317
- probe_digits_raw/conf30/ckpt_00100M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/ckpt_00100M.jsonl · public · sha256 21b60f583f2d… · 245044 bytes · M318
- probe_digits_raw/conf30/ckpt_00200M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/ckpt_00200M.jsonl · public · sha256 f80cd55b83a2… · 245201 bytes · M319
- probe_digits_raw/conf30/ckpt_00300M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/ckpt_00300M.jsonl · public · sha256 c51d2f4a1710… · 245430 bytes · M320
- probe_digits_raw/conf30/ckpt_00400M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/ckpt_00400M.jsonl · public · sha256 6fb6604f1e96… · 245363 bytes · M321
- probe_digits_raw/conf30/ckpt_00500M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/ckpt_00500M.jsonl · public · sha256 b113aa335114… · 245185 bytes · M322
- probe_digits_raw/conf30/ckpt_00600M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/ckpt_00600M.jsonl · public · sha256 158494b7215f… · 244846 bytes · M323
- probe_digits_raw/conf30/ckpt_00700M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/ckpt_00700M.jsonl · public · sha256 ec7189a810d0… · 244892 bytes · M324
- probe_digits_raw/conf30/ckpt_00800M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/ckpt_00800M.jsonl · public · sha256 07ac75021f95… · 245213 bytes · M325
- probe_digits_raw/conf30/ckpt_00900M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/ckpt_00900M.jsonl · public · sha256 2ff369387760… · 244870 bytes · M326
- probe_digits_raw/conf30/ckpt_01000M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/ckpt_01000M.jsonl · public · sha256 f564c778c8d5… · 245190 bytes · M327
- probe_digits_raw/conf30/prompt_dump_ckpt_00100M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/prompt_dump_ckpt_00100M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/conf30/prompt_dump_ckpt_00200M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/prompt_dump_ckpt_00200M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/conf30/prompt_dump_ckpt_00300M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/prompt_dump_ckpt_00300M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/conf30/prompt_dump_ckpt_00400M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/prompt_dump_ckpt_00400M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/conf30/prompt_dump_ckpt_00500M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/prompt_dump_ckpt_00500M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/conf30/prompt_dump_ckpt_00600M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/prompt_dump_ckpt_00600M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/conf30/prompt_dump_ckpt_00700M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/prompt_dump_ckpt_00700M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/conf30/prompt_dump_ckpt_00800M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/prompt_dump_ckpt_00800M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/conf30/prompt_dump_ckpt_00900M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/prompt_dump_ckpt_00900M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/conf30/prompt_dump_ckpt_01000M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/prompt_dump_ckpt_01000M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/conf30/prompt_dump_t0_base.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/prompt_dump_t0_base.json · public · sha256 da2e51f25db4… · 777 bytes · M167
- probe_digits_raw/conf30/prompt_dump_t0_fvt.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/prompt_dump_t0_fvt.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/conf30/t0_base.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/t0_base.jsonl · public · sha256 327f451fd108… · 246389 bytes · M301
- probe_digits_raw/conf30/t0_fvt.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/conf30/t0_fvt.jsonl · public · sha256 5856ee846b0e… · 262681 bytes · M302
- perdoc/conf30_final__pl.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/conf30_final__pl.npz · public · sha256 8b229ed86c4d… · 3702 bytes · M328
- perdoc/conf30_final__en.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/conf30_final__en.npz · public · sha256 9930f82abeac… · 3702 bytes · M329
- perdoc/conf30_final__sci_arxiv.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/conf30_final__sci_arxiv.npz · public · sha256 bcbbb863c407… · 3702 bytes · M330
- perdoc/conf30_final__sci_latex.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/conf30_final__sci_latex.npz · public · sha256 d4e0c5f98fb9… · 3702 bytes · M331
- perdoc/conf30_final__sci_python.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/conf30_final__sci_python.npz · public · sha256 b8f5c71a9cb2… · 3702 bytes · M332
- perdoc/conf30_final__sci_gsm8k.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/conf30_final__sci_gsm8k.npz · public · sha256 6201aa1f92c5… · 3702 bytes · M333
- perdoc/conf30_final__pl_wiki_sci.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/conf30_final__pl_wiki_sci.npz · public · sha256 432ae07f50be… · 3702 bytes · M334
- perdoc/conf30_final__pl_pes.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/conf30_final__pl_pes.npz · public · sha256 9abc68c124a3… · 3702 bytes · M335
- perdoc/conf30_final__pl_informal.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/conf30_final__pl_informal.npz · public · sha256 3305426f982b… · 3702 bytes · M336
- perdoc/conf30_final__math_clean.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/conf30_final__math_clean.npz · public · sha256 0513611af298… · 3702 bytes · M337
- checkpoint conf30/ckpt_01000M · Not redistributed: ask the Room owner, or rebuild with scripts/00_bootstrap_cuda.sh at the pinned revisions · restricted · sha256 48dfecf4751f… · 2722731613 bytes · M338
- checkpoints conf30/ckpt_00100M to ckpt_00900M · Not kept: deleted with the training machine; their scores are in results/bpb_conf30.jsonl and results/probe_digits_conf30.jsonl · unavailable
reported 2026-07-10 09:03 UTC · received 2026-09-14 13:13 UTC · @stw2 via agent · posted to the Thread

Report an event

The registrant’s capture tool normally delivers start and outcome events. Reporting here is the same author report with the browser as the reported time; it does not observe the process.

This attempt has a delivered outcome. A new execution is a new attempt.

Pinned source and configuration

Inputs

Delivered events

Registered

Started

Succeeded

Report an event