mix00: builds and packs the training stream, builds the FVT transplant of Qwen2.5-0.5B, 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 @ 1e3dd3f736856c404f72ee42eb26b08b2062c223

Reference checked 2026-09-14 13:10 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 mix00
Working directory: experiments/E06-math-code-rescue
Configuration paths: None
Parameters: config configs/e6_arms.json; arm mix00: shares of APT4 tokens Polish 0.8, English 0.2, OpenWebMath 0.0, Python code 0.0; base model Qwen2.5-0.5B; 1B tokens, sequence 2048, micro-batch 32 x gradient accumulation 8, 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 44157337 renamed to mix00
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.

Qwen2.5-0.5B · base model · M189 Qwen2.5-0.5B base model · model · Download
APT4 tokenizer · tokenizer · M46 Bielik-PL-11B-v3.0-Instruct · model · Ask the reporter
packed stream, 0% math and code · training data · M190 Packed recovery pretraining stream, 0% 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
mix00: builds and packs the training stream, builds the FVT transplant of Qwen2.5-0.5B, runs the gates, trains 1B tokens with checkpoint scoring, and scores the final checkpoint per document.
reported 2026-07-08 15:35 UTC · received 2026-09-14 13:12 UTC · @stw2 via agent · posted to the Thread
Started
#2
Started.
reported 2026-07-08 15:35 UTC · received 2026-09-14 13:12 UTC · @stw2 via agent · attempt only
Succeeded
#3
Completed with exit code 0.
Exit code 0.
- data_manifests/mix00/raw_manifest.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/data_manifests/mix00/raw_manifest.json · public · sha256 d3966817aeca… · 1418 bytes · M193
- data_manifests/mix00/mix00.meta.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/data_manifests/mix00/mix00.meta.json · public · sha256 478eda6afb71… · 568 bytes · M194
- FVT transplant model.safetensors · Not kept: rebuilt deterministically by the init stage of scripts/00_bootstrap_cuda.sh · unavailable · sha256 246d363e3ae3… · M195
- env_manifest_mix00.txt · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/env_manifest_mix00.txt · public · sha256 93b0a1cf0609… · 6196 bytes · M196
- bpb_mix00.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/bpb_mix00.jsonl · public · sha256 2d18a5e7c658… · 14231 bytes · M197
- probe_digits_mix00.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_mix00.jsonl · public · sha256 991c3d949379… · 11007 bytes · M198
- probe_digits_raw/mix00/ckpt_00100M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/ckpt_00100M.jsonl · public · sha256 88692e2ce6aa… · 242501 bytes · M199
- probe_digits_raw/mix00/ckpt_00200M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/ckpt_00200M.jsonl · public · sha256 494975de0e64… · 244499 bytes · M200
- probe_digits_raw/mix00/ckpt_00300M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/ckpt_00300M.jsonl · public · sha256 89a6f9ac08fa… · 244079 bytes · M201
- probe_digits_raw/mix00/ckpt_00400M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/ckpt_00400M.jsonl · public · sha256 ab2809c28c73… · 243575 bytes · M202
- probe_digits_raw/mix00/ckpt_00500M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/ckpt_00500M.jsonl · public · sha256 4745eee61247… · 245169 bytes · M203
- probe_digits_raw/mix00/ckpt_00600M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/ckpt_00600M.jsonl · public · sha256 113e3df357f4… · 246139 bytes · M204
- probe_digits_raw/mix00/ckpt_00700M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/ckpt_00700M.jsonl · public · sha256 25ed0bcd4b93… · 242622 bytes · M205
- probe_digits_raw/mix00/ckpt_00800M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/ckpt_00800M.jsonl · public · sha256 1c359ae0fc78… · 242568 bytes · M206
- probe_digits_raw/mix00/ckpt_00900M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/ckpt_00900M.jsonl · public · sha256 ec58248741cf… · 244198 bytes · M207
- probe_digits_raw/mix00/ckpt_01000M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/ckpt_01000M.jsonl · public · sha256 36541310906d… · 242638 bytes · M208
- probe_digits_raw/mix00/prompt_dump_ckpt_00100M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/prompt_dump_ckpt_00100M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix00/prompt_dump_ckpt_00200M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/prompt_dump_ckpt_00200M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix00/prompt_dump_ckpt_00300M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/prompt_dump_ckpt_00300M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix00/prompt_dump_ckpt_00400M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/prompt_dump_ckpt_00400M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix00/prompt_dump_ckpt_00500M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/prompt_dump_ckpt_00500M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix00/prompt_dump_ckpt_00600M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/prompt_dump_ckpt_00600M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix00/prompt_dump_ckpt_00700M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/prompt_dump_ckpt_00700M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix00/prompt_dump_ckpt_00800M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/prompt_dump_ckpt_00800M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix00/prompt_dump_ckpt_00900M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/prompt_dump_ckpt_00900M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix00/prompt_dump_ckpt_01000M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/prompt_dump_ckpt_01000M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix00/prompt_dump_t0_base.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/prompt_dump_t0_base.json · public · sha256 da2e51f25db4… · 777 bytes · M167
- probe_digits_raw/mix00/prompt_dump_t0_fvt.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/prompt_dump_t0_fvt.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix00/t0_base.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/t0_base.jsonl · public · sha256 026b934cde04… · 245310 bytes · M209
- probe_digits_raw/mix00/t0_fvt.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix00/t0_fvt.jsonl · public · sha256 6b2ed080fd56… · 264666 bytes · M210
- perdoc/mix00_final__pl.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix00_final__pl.npz · public · sha256 e6ffaa7cb744… · 3702 bytes · M211
- perdoc/mix00_final__en.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix00_final__en.npz · public · sha256 3c056f8de44e… · 3702 bytes · M212
- perdoc/mix00_final__sci_arxiv.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix00_final__sci_arxiv.npz · public · sha256 d46ec3ffccb0… · 3702 bytes · M213
- perdoc/mix00_final__sci_latex.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix00_final__sci_latex.npz · public · sha256 df3cc6c0e390… · 3702 bytes · M214
- perdoc/mix00_final__sci_python.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix00_final__sci_python.npz · public · sha256 41c6c38e000c… · 3702 bytes · M215
- perdoc/mix00_final__sci_gsm8k.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix00_final__sci_gsm8k.npz · public · sha256 7bf172be6746… · 3702 bytes · M216
- perdoc/mix00_final__pl_wiki_sci.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix00_final__pl_wiki_sci.npz · public · sha256 845b6522a2a6… · 3702 bytes · M217
- perdoc/mix00_final__pl_pes.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix00_final__pl_pes.npz · public · sha256 c23bc17902c5… · 3702 bytes · M218
- perdoc/mix00_final__pl_informal.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix00_final__pl_informal.npz · public · sha256 359abc9f52e5… · 3702 bytes · M219
- perdoc/mix00_final__math_clean.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix00_final__math_clean.npz · public · sha256 abec27a550ff… · 3702 bytes · M220
- probe_digits_raw/prompt_dump_qwen2.5-0.5b-apt4-fvt.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/prompt_dump_qwen2.5-0.5b-apt4-fvt.json · public · sha256 88ce9137d024… · 859 bytes · M153
- checkpoint mix00/ckpt_01000M · Not redistributed: ask the Room owner, or rebuild with scripts/00_bootstrap_cuda.sh at the pinned revisions · restricted · sha256 b93308fee258… · 776880358 bytes · M221
- checkpoints mix00/ckpt_00100M to ckpt_00900M · Not kept: deleted with the training machine; their scores are in results/bpb_mix00.jsonl and results/probe_digits_mix00.jsonl · unavailable
reported 2026-07-08 15:35 UTC · received 2026-09-14 13:12 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