Delower's picture
Upload folder using huggingface_hub
ea77201 verified
Raw History Blame Contribute Delete
3.79 kB
{
"project": "hERG toxicity prediction - ExtraTrees-comprehensive DEPLOYMENT REFITS (third deployment model; Phase-5 best single model, refit on all labeled data)",
"version": "deployment-refit-2026-11-16",
"deployment_refit": true,
"trained_on": "ALL labeled data: locked pools (cls 12,702 / reg 8,657) + masters remainder + absorbed external sets (hERG-70 462, new-compounds 967, FDA panel 36, REGFAIL 7, hERG-60, Test-1)",
"training_files": {
"classification": "cls_alldata.csv (14,441 rows)",
"regression": "reg_alldata.csv (10,370 rows)"
},
"hard_gate": "every external/regulatory compound verified present in the actual training files AND in the fitted training matrices by InChIKey connectivity (counts in validation_results.json; full report HARD_GATE_ET.md in the project record)",
"models": {
"cls_scaffold": {
"file": "models/extratreesA_cls_morgan+maccs+desc+physchem_extra+herg+pharm_scaffold_DEPLOYMENT.pkl",
"task": "classification",
"protocol": "scaffold",
"headline": true,
"n_train": 14441,
"n_features": 2499
},
"cls_random": {
"file": "models/extratreesA_cls_morgan+maccs+desc+physchem_extra+herg+pharm_random_DEPLOYMENT.pkl",
"task": "classification",
"protocol": "random",
"headline": false,
"n_train": 14441,
"n_features": 2499
},
"reg_scaffold": {
"file": "models/extratreesA_reg_morgan+desc+herg+pharm_scaffold_DEPLOYMENT.pkl",
"task": "regression",
"protocol": "scaffold",
"headline": true,
"n_train": 10370,
"n_features": 2331
},
"reg_random": {
"file": "models/extratreesA_reg_morgan+desc+herg+pharm_random_DEPLOYMENT.pkl",
"task": "regression",
"protocol": "random",
"headline": false,
"n_train": 10370,
"n_features": 2331
}
},
"feature_blocks": {
"classification": [
"morgan",
"maccs",
"desc",
"physchem_extra",
"herg",
"pharm"
],
"regression": [
"morgan",
"desc",
"herg",
"pharm"
]
},
"hyperparameters": {
"n_estimators": 300,
"n_jobs": 2,
"random_state": 42,
"note": "identical to the locked-pool Phase-5 ET finals (p5_approachA_et.py)"
},
"protocol_note": "scaffold and random bundles wrap the same fitted model per task: identical combo/hyperparameters and all-data training make the protocol distinction vanish in deployment refits",
"honest_locked_benchmark_scores": {
"source": "Locked-pool Phase-5 refits (train+val only; every external/regulatory set locked out), results/phase5/PHASE5_RESULTS.md \u2014 these are the honest benchmark scores of this architecture. The DEPLOYMENT refits in this repo absorbed those sets and are never scored on them.",
"classification_scaffold_refit": {
"hERG-70": {
"ACC": 0.76,
"MCC": 0.543
},
"new_compounds_MCC": 0.149,
"FDA_errors": "4/36",
"REGFAIL_Tier1": "4/5"
},
"classification_random_refit": {
"hERG-70": {
"ACC": 0.74,
"MCC": 0.504
},
"new_compounds_MCC": 0.154,
"FDA_errors": "5/36",
"REGFAIL_Tier1": "4/5"
},
"regression_scaffold_refit": {
"hERG-70_RMSE": 0.838
},
"regression_random_refit": {
"hERG-70_RMSE": 0.829
},
"note": "Best SINGLE (non-fusion) model on hERG-70 in the programme (MCC 0.543); the best overall classifier is Fusion G (0.548, XGBoost + CardioGenAI), which is not a deployment model."
},
"sibling_deployment_repos": [
"Delower/herg-xgboost-deployment",
"Delower/herg-cardiogenai-deployment"
],
"warning": "DEPLOYMENT MODEL - trained on the regulatory/external sets. Do NOT report its scores on those sets as validation."
}