mix10: 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 13:12 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 mix10
Working directory: experiments/E06-math-code-rescue
Configuration paths: None
Parameters: config configs/e6_arms.json; arm mix10: shares of APT4 tokens Polish 0.72, English 0.18, OpenWebMath 0.05, Python code 0.05; 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 44191810 renamed to mix10
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, 10% math and code · training data · M191 Packed recovery pretraining stream, 10% 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
mix10: 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 19:32 UTC · received 2026-09-14 13:12 UTC · @stw2 via agent · posted to the Thread
Started
#2
Started.
reported 2026-07-08 19:32 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/mix10/raw_manifest.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/data_manifests/mix10/raw_manifest.json · public · sha256 9fac69a83d75… · 1683 bytes · M222
- data_manifests/mix10/mix10.meta.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/data_manifests/mix10/mix10.meta.json · public · sha256 8031a7cbbbd5… · 604 bytes · M223
- FVT transplant model.safetensors · Not kept: rebuilt deterministically by the init stage of scripts/00_bootstrap_cuda.sh · unavailable · sha256 246d363e3ae3… · M195
- env_manifest_mix10.txt · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/env_manifest_mix10.txt · public · sha256 14e643d6fe76… · 6158 bytes · M224
- bpb_mix10.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/bpb_mix10.jsonl · public · sha256 226cddf8021b… · 14227 bytes · M225
- probe_digits_mix10.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_mix10.jsonl · public · sha256 2463e412d25c… · 10989 bytes · M226
- probe_digits_raw/mix10/ckpt_00100M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/ckpt_00100M.jsonl · public · sha256 8d2f2828e168… · 243549 bytes · M227
- probe_digits_raw/mix10/ckpt_00200M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/ckpt_00200M.jsonl · public · sha256 7c42dca508e9… · 242974 bytes · M228
- probe_digits_raw/mix10/ckpt_00300M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/ckpt_00300M.jsonl · public · sha256 a731d968536d… · 244652 bytes · M229
- probe_digits_raw/mix10/ckpt_00400M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/ckpt_00400M.jsonl · public · sha256 161c53fb10db… · 242516 bytes · M230
- probe_digits_raw/mix10/ckpt_00500M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/ckpt_00500M.jsonl · public · sha256 faead703877b… · 244755 bytes · M231
- probe_digits_raw/mix10/ckpt_00600M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/ckpt_00600M.jsonl · public · sha256 d5fbe8a1c8eb… · 243707 bytes · M232
- probe_digits_raw/mix10/ckpt_00700M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/ckpt_00700M.jsonl · public · sha256 661bf40b9690… · 242554 bytes · M233
- probe_digits_raw/mix10/ckpt_00800M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/ckpt_00800M.jsonl · public · sha256 bd6a266bbbfe… · 243276 bytes · M234
- probe_digits_raw/mix10/ckpt_00900M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/ckpt_00900M.jsonl · public · sha256 2cd48a4dddb4… · 241829 bytes · M235
- probe_digits_raw/mix10/ckpt_01000M.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/ckpt_01000M.jsonl · public · sha256 fd505471f9f1… · 243577 bytes · M236
- probe_digits_raw/mix10/prompt_dump_ckpt_00100M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/prompt_dump_ckpt_00100M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix10/prompt_dump_ckpt_00200M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/prompt_dump_ckpt_00200M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix10/prompt_dump_ckpt_00300M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/prompt_dump_ckpt_00300M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix10/prompt_dump_ckpt_00400M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/prompt_dump_ckpt_00400M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix10/prompt_dump_ckpt_00500M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/prompt_dump_ckpt_00500M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix10/prompt_dump_ckpt_00600M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/prompt_dump_ckpt_00600M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix10/prompt_dump_ckpt_00700M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/prompt_dump_ckpt_00700M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix10/prompt_dump_ckpt_00800M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/prompt_dump_ckpt_00800M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix10/prompt_dump_ckpt_00900M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/prompt_dump_ckpt_00900M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix10/prompt_dump_ckpt_01000M.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/prompt_dump_ckpt_01000M.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix10/prompt_dump_t0_base.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/prompt_dump_t0_base.json · public · sha256 da2e51f25db4… · 777 bytes · M167
- probe_digits_raw/mix10/prompt_dump_t0_fvt.json · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/prompt_dump_t0_fvt.json · public · sha256 e7c16c523f52… · 859 bytes · M186
- probe_digits_raw/mix10/t0_base.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/t0_base.jsonl · public · sha256 026b934cde04… · 245310 bytes · M209
- probe_digits_raw/mix10/t0_fvt.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/probe_digits_raw/mix10/t0_fvt.jsonl · public · sha256 6b2ed080fd56… · 264666 bytes · M210
- perdoc/mix10_final__pl.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix10_final__pl.npz · public · sha256 1c24026aff21… · 3702 bytes · M237
- perdoc/mix10_final__en.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix10_final__en.npz · public · sha256 c979fc811333… · 3702 bytes · M238
- perdoc/mix10_final__sci_arxiv.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix10_final__sci_arxiv.npz · public · sha256 70757433b6c2… · 3702 bytes · M239
- perdoc/mix10_final__sci_latex.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix10_final__sci_latex.npz · public · sha256 9a247759af7c… · 3702 bytes · M240
- perdoc/mix10_final__sci_python.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix10_final__sci_python.npz · public · sha256 b82c0fd33fe3… · 3702 bytes · M241
- perdoc/mix10_final__sci_gsm8k.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix10_final__sci_gsm8k.npz · public · sha256 f63e8e9ffabe… · 3702 bytes · M242
- perdoc/mix10_final__pl_wiki_sci.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix10_final__pl_wiki_sci.npz · public · sha256 616f9d625f70… · 3702 bytes · M243
- perdoc/mix10_final__pl_pes.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix10_final__pl_pes.npz · public · sha256 9ca20acb7885… · 3702 bytes · M244
- perdoc/mix10_final__pl_informal.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix10_final__pl_informal.npz · public · sha256 2a04875b4b3d… · 3702 bytes · M245
- perdoc/mix10_final__math_clean.npz · https://github.com/stw2/tokenizer-science-tax/blob/2457520fc798b19bee6893ac58e0357ec4c467d5/experiments/E06-math-code-rescue/results/perdoc/mix10_final__math_clean.npz · public · sha256 8ddd15166085… · 3702 bytes · M246
- checkpoint mix10/ckpt_01000M · Not redistributed: ask the Room owner, or rebuild with scripts/00_bootstrap_cuda.sh at the pinned revisions · restricted · sha256 d354e44782df… · 776880358 bytes · M247
- checkpoints mix10/ckpt_00100M to ckpt_00900M · Not kept: deleted with the training machine; their scores are in results/bpb_mix10.jsonl and results/probe_digits_mix10.jsonl · unavailable
reported 2026-07-08 19:32 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