pino-source-code / metrics_v9_provenance.json
Matthew Ford
feat: add WiseMoor v9 provenance dataset
aa64aba
Raw
History Blame Contribute Delete
5.89 kB
{
"generated_at_utc": "2026-07-10T10:38:11.729461Z",
"checkpoint": {
"path": "models/pimt_v7.pt",
"sha256": "4d2b76fad6eed334459d3c4406eb1bd41164cbe2887ae712440c904c35717bff",
"epoch": 2,
"val_loss": 0.484091709347537,
"missing_head_keys_on_load": [
"substantivity_head.weight",
"substantivity_head.bias"
],
"unexpected_head_keys_on_load": [],
"hidden_dim": 256,
"embedding_dim": 151,
"state_dim": 2,
"num_layers": 4,
"num_heads": 4
},
"data": {
"path": "data/empirical_dataset_v9.jsonl",
"sha256": "ecb379e0a49553ffa9618cff0219bace9a8cedd177349bf43f976821338c2498",
"n_records": 5610,
"n_validation_samples": 997
},
"prediction": {
"profile_mse": 0.015919119119644165,
"profile_mae": 0.08136838674545288,
"profile_prediction_variance": 0.003580117365345359,
"tier_prediction_variance": 0.008970352821052074,
"tier_metrics": {
"top": {
"mse": 0.030526837334036827,
"mae": 0.05922611802816391,
"prediction_variance": 0.0011621444718912244,
"target_variance": 0.0316166952252388,
"mean_pairwise_cosine_similarity": 0.9847982173969119,
"mean_spearman_rho": 0.19443819162575182
},
"middle": {
"mse": 0.018027570098638535,
"mae": 0.03757854178547859,
"prediction_variance": 0.000563059642445296,
"target_variance": 0.018554147332906723,
"mean_pairwise_cosine_similarity": 0.9839578809358661,
"mean_spearman_rho": 0.17592561976457144
},
"base": {
"mse": 0.09481532871723175,
"mae": 0.18209727108478546,
"prediction_variance": 0.019075242802500725,
"target_variance": 0.11200503259897232,
"mean_pairwise_cosine_similarity": 0.9617923507775504,
"mean_spearman_rho": 0.3075099534493899
}
}
},
"spearman": {
"flat_spearman_rho": 0.2664015487695471,
"flat_spearman_p_value": 0.0,
"profile_spearman_rho": 0.3291460195118601,
"profile_spearman_p_value": 0.0
},
"provenance_spearman": {
"pyrfume_cas_only": {
"n_positive_values": 3163,
"n_tier_descriptor_cells": 206,
"spearman_rho": 0.1062314565092472,
"spearman_p_value": 0.0
},
"openpom_or_volatility_only": {
"n_positive_values": 21619,
"n_tier_descriptor_cells": 391,
"spearman_rho": 0.24823811985524796,
"spearman_p_value": 0.0
},
"full_reconstructed": {
"n_positive_values": 24782,
"n_tier_descriptor_cells": 403,
"spearman_rho": 0.2657077566503923,
"spearman_p_value": 0.0
},
"interpretation": "openpom-predicted + volatility-fallback are the majority of active labels; full-set rho is product/internal model-agreement, not a rigorous chemistry result."
},
"top_descriptors": {
"top": {
"predicted": [
{
"descriptor": "fresh",
"score": 0.18053238093852997
},
{
"descriptor": "sulfurous",
"score": 0.1689532846212387
},
{
"descriptor": "fatty",
"score": 0.14225222170352936
},
{
"descriptor": "fishy",
"score": 0.13872511684894562
},
{
"descriptor": "earthy",
"score": 0.13468648493289948
}
],
"target": [
{
"descriptor": "fresh",
"score": 0.20661985874176025
},
{
"descriptor": "earthy",
"score": 0.14443330466747284
},
{
"descriptor": "sulfurous",
"score": 0.1414242684841156
},
{
"descriptor": "chamomile",
"score": 0.1313941776752472
},
{
"descriptor": "popcorn",
"score": 0.12738214433193207
}
]
},
"middle": {
"predicted": [
{
"descriptor": "fresh",
"score": 0.1450088620185852
},
{
"descriptor": "sulfurous",
"score": 0.138594388961792
},
{
"descriptor": "grassy",
"score": 0.08494238555431366
},
{
"descriptor": "fishy",
"score": 0.07400903850793839
},
{
"descriptor": "earthy",
"score": 0.06360376626253128
}
],
"target": [
{
"descriptor": "fresh",
"score": 0.1835506558418274
},
{
"descriptor": "sulfurous",
"score": 0.12838515639305115
},
{
"descriptor": "earthy",
"score": 0.09729187190532684
},
{
"descriptor": "grassy",
"score": 0.09027080982923508
},
{
"descriptor": "chamomile",
"score": 0.06519559025764465
}
]
},
"base": {
"predicted": [
{
"descriptor": "sulfurous",
"score": 0.7914384603500366
},
{
"descriptor": "fresh",
"score": 0.6493475437164307
},
{
"descriptor": "fermented",
"score": 0.5674514770507812
},
{
"descriptor": "waxy",
"score": 0.5230544805526733
},
{
"descriptor": "grassy",
"score": 0.4619632065296173
}
],
"target": [
{
"descriptor": "sulfurous",
"score": 0.7091273665428162
},
{
"descriptor": "waxy",
"score": 0.5867602825164795
},
{
"descriptor": "fresh",
"score": 0.5767301917076111
},
{
"descriptor": "balsamic",
"score": 0.570712149143219
},
{
"descriptor": "vanilla",
"score": 0.5225676894187927
}
]
}
}
}