Automatic search from v0.45: adds who funded the early buyers, the creator's wallet and real on-chain holders, plus every newer signal; min samples leaf 40 → 5, with automatic signal discovery.
Failed
Definition
Model type
Boosted decision trees
Signals used
… + funding links & on-chain holders
Technical details
max_depth
3
subsample
0.8
n_estimators
300
learning_rate
0.03
min_samples_leaf
5
Random seed
1056
Config hash
79ca114babf8cd37
Created
12 Oct 08:02
Completed
12 Oct 08:02
Artifact
—
Dataset & compute
Dataset
Solana Launch Dataset v39 (f10)
66 snapshots · 14 launches · sha256 0049fcfa614a
Training runs
0
Compute
—
GPU hours
—
no GPU used
CPU time
0.0 s
Provider cost
$0.00
Benchmarks
held-out test set · change vs v0.46
No benchmark numbers published yet — the locked test set needs more held-out launches with observed outcomes.
What this version relies on
held-out validation launches
This version was trained before Model Zero started measuring which signals each model relies on.
For each signal, how much the score drops when that signal is scrambled across launches. Model Zero learns these weights itself from the data; nobody hand-codes them.
Training runs
No run yet.
Timeline
12 Oct 08:02
Model Zero v0.56 not trained
dataset too small: 66 snapshots on f10 against 168,694 for the live v0.45; that feature set covers too few launches yet.
12 Oct 08:02
Model Zero v0.56 queued
Automatic search from v0.45: adds who funded the early buyers, the creator's wallet and real on-chain holders, plus every newer signal; min samples leaf 40 → 5, with automatic signal discovery.
12 Oct 08:02
Enough new data for another training run
224,252 new labelled snapshots since Solana Launch Dataset v9.