Buckets:
Hugging Face Jobs audit trail
All runs used the authenticated SabaPivot namespace on 2026-07-21.
| Purpose | Job | Hardware | Runtime | Result |
|---|---|---|---|---|
| Infrastructure canary | https://huggingface.co/jobs/SabaPivot/6a5f1ee413e6ef894d549757 | CPU basic | 0 s | completed |
| Core claims, 64 seeds | https://huggingface.co/jobs/SabaPivot/6a5f2044d09dc1f57c6bed22 | Tesla T4 | 41 s | completed; /results/core persisted |
| Architecture spectra, first attempt | https://huggingface.co/jobs/SabaPivot/6a5f205113e6ef894d549760 | Tesla T4 | 63 s | inference completed; write failed after bucket leaf refresh |
| Architecture spectra, corrected retry | https://huggingface.co/jobs/SabaPivot/6a5f2198d09dc1f57c6bed3c | Tesla T4 | 59 s | completed; /results/architecture persisted |
Core command:
uv run --with numpy --with torch /data/core_repro.py --output /results/core --reps 64 --seed 23301 --backend torch
Architecture command:
uv run /data/architecture_spectra.py --output /results/architecture --samples 1000 --batch-size 32
Results: https://huggingface.co/buckets/SabaPivot/repro-what-makes-a-strong-model-results
Staged job inputs: https://huggingface.co/buckets/SabaPivot/jobs-artifacts
Xet Storage Details
- Size:
- 1.26 kB
- Xet hash:
- 4e37f1d222f1961dcfebb59131f1b3fb2ab9771fe909ec8aeb37ae54fa5bf524
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.