Downloads · 30 days
2
4% of all-time downloads
tope1129/cas-smo-v30
cas-smo-v30 is a machine learning model from tope1129. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Pool: dataset/snapshots/2026-08-04-block-8769600 (1097 series; ~30% transport / 22% energy / 20% nature).
Downloads · 30 days
2
4% of all-time downloads
All-time downloads
51
Public
Repo size
—
Likes
0
Public
Click a slice to open those files.
.py68.4 KB · 90%
From the Hugging Face model README
Pool: dataset/snapshots/2026-08-04-block-8769600 (1097 series; ~30% transport / 22% energy / 20% nature).
| Generator | geomean (seed 0, heat 0.5h / 256 win) |
|---|---|
| beyondyourself113/cascade-y3 | 0.20329 |
| angry-bird/gen-d79b2f553e11 | 0.19661 |
| cas-smo-v30 (= v28) | 0.19353 |
| Target (best top-2 − 0.01) | 0.18661 |
Beats both current top-2 by 0.00308 vs angry-bird (not the full 0.01).
Promotion of cas-smo-v28 (step jump amp uniform(0.5, 2.5), exact 0.78, integral 0.82; family mix with step_level 0.45 + conditional_stability 0.21 + ar2 0.235 + small epi/grid/flow).
Tried and rejected on this pool (full heat unless noted): angry+v28-step, large specialist rebalance, sticky 0.06, nonneg prior, sticky 0.02 (0.19417), sticky+grid+tidal (0.19566). Short-screen sticky/grid leads did not survive full confirm.