mirror of
https://github.com/RYDE-WORK/lnp_ml.git
synced 2026-07-23 06:11:37 +08:00
977 lines
28 KiB
JSON
977 lines
28 KiB
JSON
{
|
|
"fold_results": [
|
|
{
|
|
"fold": 0,
|
|
"best_params": {
|
|
"dropout": 0.11659837578806788,
|
|
"lr": 0.00026604612647882186,
|
|
"weight_decay": 0.00021355962098733124,
|
|
"backbone_lr_ratio": 0.29733721720121176,
|
|
"moe_n_experts": 2,
|
|
"moe_top_k": 2,
|
|
"moe_expert_hidden_mult": 2,
|
|
"d_model": 256,
|
|
"num_heads": 8,
|
|
"n_attn_layers": 4,
|
|
"fusion_strategy": "attention",
|
|
"head_hidden_dim": 128,
|
|
"set_transformer_block": "sab"
|
|
},
|
|
"epoch_mean": 18,
|
|
"test_metrics": {
|
|
"size": {
|
|
"n_samples": 83,
|
|
"mse": 0.24151334347557607,
|
|
"rmse": 0.4914400710926777,
|
|
"mae": 0.25384363496159934,
|
|
"r2": 0.1980890419336545
|
|
},
|
|
"delivery": {
|
|
"n_samples": 58,
|
|
"mse": 0.6215165843621644,
|
|
"rmse": 0.7883632312342861,
|
|
"mae": 0.6303568019949156,
|
|
"r2": 0.21001491653032867
|
|
},
|
|
"pdi": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.5833333333333334,
|
|
"precision": 0.34911616161616166,
|
|
"recall": 0.532591725214676,
|
|
"f1": 0.3726817042606517
|
|
},
|
|
"ee": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.7261904761904762,
|
|
"precision": 0.6833403612340994,
|
|
"recall": 0.730952380952381,
|
|
"f1": 0.6870884387763607
|
|
},
|
|
"toxic": {
|
|
"n_samples": 58,
|
|
"accuracy": 0.9655172413793104,
|
|
"precision": 0.75,
|
|
"recall": 0.9821428571428572,
|
|
"f1": 0.8242424242424242
|
|
},
|
|
"biodist": {
|
|
"n_samples": 58,
|
|
"kl_divergence": 0.3475976565313305,
|
|
"js_divergence": 0.08236128812465786
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"fold": 1,
|
|
"best_params": {
|
|
"dropout": 0.18614946459458093,
|
|
"lr": 0.00015757063904422037,
|
|
"weight_decay": 0.000955120409040689,
|
|
"backbone_lr_ratio": 0.9059364477738808,
|
|
"moe_n_experts": 4,
|
|
"moe_top_k": 2,
|
|
"moe_expert_hidden_mult": 1,
|
|
"d_model": 256,
|
|
"num_heads": 8,
|
|
"n_attn_layers": 4,
|
|
"fusion_strategy": "attention",
|
|
"head_hidden_dim": 128,
|
|
"set_transformer_block": "sab"
|
|
},
|
|
"epoch_mean": 16,
|
|
"test_metrics": {
|
|
"size": {
|
|
"n_samples": 84,
|
|
"mse": 0.07934253557006275,
|
|
"rmse": 0.281678070800804,
|
|
"mae": 0.21441291059766496,
|
|
"r2": -0.3144397304964943
|
|
},
|
|
"delivery": {
|
|
"n_samples": 61,
|
|
"mse": 1.0472045646776351,
|
|
"rmse": 1.0233301347452028,
|
|
"mae": 0.6786254059034781,
|
|
"r2": 0.20734163639743197
|
|
},
|
|
"pdi": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.6309523809523809,
|
|
"precision": 0.2611111111111111,
|
|
"recall": 0.24206349206349206,
|
|
"f1": 0.25
|
|
},
|
|
"ee": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.5595238095238095,
|
|
"precision": 0.49978354978354983,
|
|
"recall": 0.49526663812378097,
|
|
"f1": 0.48344031447479724
|
|
},
|
|
"toxic": {
|
|
"n_samples": 61,
|
|
"accuracy": 0.9508196721311475,
|
|
"precision": 0.75,
|
|
"recall": 0.9741379310344828,
|
|
"f1": 0.8200589970501475
|
|
},
|
|
"biodist": {
|
|
"n_samples": 61,
|
|
"kl_divergence": 0.4763434813369227,
|
|
"js_divergence": 0.12345390425610421
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"fold": 2,
|
|
"best_params": {
|
|
"dropout": 0.27066334445695367,
|
|
"lr": 0.00024715599869840884,
|
|
"weight_decay": 1.0172836375002765e-05,
|
|
"backbone_lr_ratio": 0.5054041560165133,
|
|
"moe_n_experts": 4,
|
|
"moe_top_k": 1,
|
|
"moe_expert_hidden_mult": 2,
|
|
"d_model": 256,
|
|
"num_heads": 8,
|
|
"n_attn_layers": 4,
|
|
"fusion_strategy": "attention",
|
|
"head_hidden_dim": 128,
|
|
"set_transformer_block": "sab"
|
|
},
|
|
"epoch_mean": 14,
|
|
"test_metrics": {
|
|
"size": {
|
|
"n_samples": 84,
|
|
"mse": 0.1393018116424732,
|
|
"rmse": 0.37323157910668975,
|
|
"mae": 0.30152987014679683,
|
|
"r2": -0.49257454331973105
|
|
},
|
|
"delivery": {
|
|
"n_samples": 60,
|
|
"mse": 0.6391425542093007,
|
|
"rmse": 0.7994639167650411,
|
|
"mae": 0.5482877640752122,
|
|
"r2": 0.14919048166659254
|
|
},
|
|
"pdi": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.5714285714285714,
|
|
"precision": 0.4258597883597883,
|
|
"recall": 0.6909882232462877,
|
|
"f1": 0.4396490918709222
|
|
},
|
|
"ee": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.6666666666666666,
|
|
"precision": 0.6053508771929824,
|
|
"recall": 0.6573295985060691,
|
|
"f1": 0.6168590642643939
|
|
},
|
|
"toxic": {
|
|
"n_samples": 61,
|
|
"accuracy": 0.9344262295081968,
|
|
"precision": 0.7142857142857143,
|
|
"recall": 0.9655172413793103,
|
|
"f1": 0.7821428571428571
|
|
},
|
|
"biodist": {
|
|
"n_samples": 60,
|
|
"kl_divergence": 0.29610571824726867,
|
|
"js_divergence": 0.07901143607137948
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"fold": 3,
|
|
"best_params": {
|
|
"dropout": 0.2778201867797741,
|
|
"lr": 0.0004781288818954282,
|
|
"weight_decay": 0.0016987970387862138,
|
|
"backbone_lr_ratio": 0.9458696393367624,
|
|
"moe_n_experts": 4,
|
|
"moe_top_k": 2,
|
|
"moe_expert_hidden_mult": 2,
|
|
"d_model": 256,
|
|
"num_heads": 8,
|
|
"n_attn_layers": 4,
|
|
"fusion_strategy": "attention",
|
|
"head_hidden_dim": 128,
|
|
"set_transformer_block": "sab"
|
|
},
|
|
"epoch_mean": 15,
|
|
"test_metrics": {
|
|
"size": {
|
|
"n_samples": 83,
|
|
"mse": 0.3258503637013872,
|
|
"rmse": 0.5708330436313118,
|
|
"mae": 0.35500740717692547,
|
|
"r2": 0.00011614402893123099
|
|
},
|
|
"delivery": {
|
|
"n_samples": 59,
|
|
"mse": 0.9100019552700145,
|
|
"rmse": 0.9539402262563491,
|
|
"mae": 0.6691959031652344,
|
|
"r2": 0.11536998132084963
|
|
},
|
|
"pdi": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.7023809523809523,
|
|
"precision": 0.3551587301587302,
|
|
"recall": 0.5311059907834101,
|
|
"f1": 0.3651428571428571
|
|
},
|
|
"ee": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.6547619047619048,
|
|
"precision": 0.5822944896115628,
|
|
"recall": 0.6315736551030668,
|
|
"f1": 0.583905415713196
|
|
},
|
|
"toxic": {
|
|
"n_samples": 60,
|
|
"accuracy": 0.9333333333333333,
|
|
"precision": 0.7142857142857143,
|
|
"recall": 0.9649122807017544,
|
|
"f1": 0.7818181818181817
|
|
},
|
|
"biodist": {
|
|
"n_samples": 60,
|
|
"kl_divergence": 0.15341130227627944,
|
|
"js_divergence": 0.03932574094762976
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"fold": 4,
|
|
"best_params": {
|
|
"dropout": 0.445253277169985,
|
|
"lr": 0.0002457483875420317,
|
|
"weight_decay": 0.00014672230276933923,
|
|
"backbone_lr_ratio": 0.2588275309142304,
|
|
"moe_n_experts": 2,
|
|
"moe_top_k": 2,
|
|
"moe_expert_hidden_mult": 2,
|
|
"d_model": 256,
|
|
"num_heads": 8,
|
|
"n_attn_layers": 4,
|
|
"fusion_strategy": "attention",
|
|
"head_hidden_dim": 128,
|
|
"set_transformer_block": "sab"
|
|
},
|
|
"epoch_mean": 25,
|
|
"test_metrics": {
|
|
"size": {
|
|
"n_samples": 84,
|
|
"mse": 0.3157862632738926,
|
|
"rmse": 0.561948630458241,
|
|
"mae": 0.46222461689086186,
|
|
"r2": -1.2924811440596238
|
|
},
|
|
"delivery": {
|
|
"n_samples": 58,
|
|
"mse": 0.894030368182473,
|
|
"rmse": 0.945531791206659,
|
|
"mae": 0.6543392465160839,
|
|
"r2": 0.1354873927120125
|
|
},
|
|
"pdi": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.5714285714285714,
|
|
"precision": 0.3313131313131313,
|
|
"recall": 0.5182411674347158,
|
|
"f1": 0.34882312218760814
|
|
},
|
|
"ee": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.6904761904761905,
|
|
"precision": 0.6414141414141414,
|
|
"recall": 0.6649572649572649,
|
|
"f1": 0.6458333333333334
|
|
},
|
|
"toxic": {
|
|
"n_samples": 59,
|
|
"accuracy": 0.9661016949152542,
|
|
"precision": 0.8,
|
|
"recall": 0.9821428571428572,
|
|
"f1": 0.8659090909090909
|
|
},
|
|
"biodist": {
|
|
"n_samples": 58,
|
|
"kl_divergence": 0.41797103371550187,
|
|
"js_divergence": 0.1075696933374844
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"fold": 0,
|
|
"best_params": {
|
|
"dropout": 0.11659837578806788,
|
|
"lr": 0.00026604612647882186,
|
|
"weight_decay": 0.00021355962098733124,
|
|
"backbone_lr_ratio": 0.29733721720121176,
|
|
"moe_n_experts": 2,
|
|
"moe_top_k": 2,
|
|
"moe_expert_hidden_mult": 2,
|
|
"d_model": 256,
|
|
"num_heads": 8,
|
|
"n_attn_layers": 4,
|
|
"fusion_strategy": "attention",
|
|
"head_hidden_dim": 128,
|
|
"set_transformer_block": "sab"
|
|
},
|
|
"epoch_mean": 18,
|
|
"test_metrics": {
|
|
"size": {
|
|
"n_samples": 83,
|
|
"mse": 0.2913344988852884,
|
|
"rmse": 0.539754109651134,
|
|
"mae": 0.2935657845922263,
|
|
"r2": 0.03266492957766376
|
|
},
|
|
"delivery": {
|
|
"n_samples": 58,
|
|
"mse": 0.6689203995885333,
|
|
"rmse": 0.8178755404024095,
|
|
"mae": 0.6310591241023664,
|
|
"r2": 0.1497618068457085
|
|
},
|
|
"pdi": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.5833333333333334,
|
|
"precision": 0.3387987012987013,
|
|
"recall": 0.532591725214676,
|
|
"f1": 0.3577380952380953
|
|
},
|
|
"ee": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.6904761904761905,
|
|
"precision": 0.688189110769756,
|
|
"recall": 0.7452380952380953,
|
|
"f1": 0.6821408529741864
|
|
},
|
|
"toxic": {
|
|
"n_samples": 58,
|
|
"accuracy": 0.9827586206896551,
|
|
"precision": 0.9912280701754386,
|
|
"recall": 0.75,
|
|
"f1": 0.8289085545722714
|
|
},
|
|
"biodist": {
|
|
"n_samples": 58,
|
|
"kl_divergence": 0.3378843364368577,
|
|
"js_divergence": 0.0825013706213184
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"fold": 1,
|
|
"best_params": {
|
|
"dropout": 0.18614946459458093,
|
|
"lr": 0.00015757063904422037,
|
|
"weight_decay": 0.000955120409040689,
|
|
"backbone_lr_ratio": 0.9059364477738808,
|
|
"moe_n_experts": 4,
|
|
"moe_top_k": 2,
|
|
"moe_expert_hidden_mult": 1,
|
|
"d_model": 256,
|
|
"num_heads": 8,
|
|
"n_attn_layers": 4,
|
|
"fusion_strategy": "attention",
|
|
"head_hidden_dim": 128,
|
|
"set_transformer_block": "sab"
|
|
},
|
|
"epoch_mean": 16,
|
|
"test_metrics": {
|
|
"size": {
|
|
"n_samples": 84,
|
|
"mse": 0.07875340556972314,
|
|
"rmse": 0.28063037178773637,
|
|
"mae": 0.2168605923652649,
|
|
"r2": -0.3046798221029694
|
|
},
|
|
"delivery": {
|
|
"n_samples": 61,
|
|
"mse": 1.0042013711545308,
|
|
"rmse": 1.0020984837602194,
|
|
"mae": 0.6381056923419237,
|
|
"r2": 0.2398919538401385
|
|
},
|
|
"pdi": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.5357142857142857,
|
|
"precision": 0.28846153846153844,
|
|
"recall": 0.27380952380952384,
|
|
"f1": 0.26176470588235295
|
|
},
|
|
"ee": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.5595238095238095,
|
|
"precision": 0.49570604833762727,
|
|
"recall": 0.49526663812378097,
|
|
"f1": 0.4855665407389545
|
|
},
|
|
"toxic": {
|
|
"n_samples": 61,
|
|
"accuracy": 0.9508196721311475,
|
|
"precision": 0.75,
|
|
"recall": 0.9741379310344828,
|
|
"f1": 0.8200589970501475
|
|
},
|
|
"biodist": {
|
|
"n_samples": 61,
|
|
"kl_divergence": 0.44548003202564046,
|
|
"js_divergence": 0.11344950900067649
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"fold": 2,
|
|
"best_params": {
|
|
"dropout": 0.27066334445695367,
|
|
"lr": 0.00024715599869840884,
|
|
"weight_decay": 1.0172836375002765e-05,
|
|
"backbone_lr_ratio": 0.5054041560165133,
|
|
"moe_n_experts": 4,
|
|
"moe_top_k": 1,
|
|
"moe_expert_hidden_mult": 2,
|
|
"d_model": 256,
|
|
"num_heads": 8,
|
|
"n_attn_layers": 4,
|
|
"fusion_strategy": "attention",
|
|
"head_hidden_dim": 128,
|
|
"set_transformer_block": "sab"
|
|
},
|
|
"epoch_mean": 14,
|
|
"test_metrics": {
|
|
"size": {
|
|
"n_samples": 84,
|
|
"mse": 0.21586856281890104,
|
|
"rmse": 0.464616576134452,
|
|
"mae": 0.3479909045355661,
|
|
"r2": -1.3129628952238908
|
|
},
|
|
"delivery": {
|
|
"n_samples": 60,
|
|
"mse": 0.653652792026344,
|
|
"rmse": 0.8084879665315644,
|
|
"mae": 0.5382411242850745,
|
|
"r2": 0.1298748401610842
|
|
},
|
|
"pdi": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.5357142857142857,
|
|
"precision": 0.4276315789473684,
|
|
"recall": 0.6958525345622121,
|
|
"f1": 0.42300062774639047
|
|
},
|
|
"ee": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.6666666666666666,
|
|
"precision": 0.6053508771929824,
|
|
"recall": 0.6573295985060691,
|
|
"f1": 0.6168590642643939
|
|
},
|
|
"toxic": {
|
|
"n_samples": 61,
|
|
"accuracy": 0.9344262295081968,
|
|
"precision": 0.7142857142857143,
|
|
"recall": 0.9655172413793103,
|
|
"f1": 0.7821428571428571
|
|
},
|
|
"biodist": {
|
|
"n_samples": 60,
|
|
"kl_divergence": 0.30110808565332015,
|
|
"js_divergence": 0.07808617669952787
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"fold": 3,
|
|
"best_params": {
|
|
"dropout": 0.2778201867797741,
|
|
"lr": 0.0004781288818954282,
|
|
"weight_decay": 0.0016987970387862138,
|
|
"backbone_lr_ratio": 0.9458696393367624,
|
|
"moe_n_experts": 4,
|
|
"moe_top_k": 2,
|
|
"moe_expert_hidden_mult": 2,
|
|
"d_model": 256,
|
|
"num_heads": 8,
|
|
"n_attn_layers": 4,
|
|
"fusion_strategy": "attention",
|
|
"head_hidden_dim": 128,
|
|
"set_transformer_block": "sab"
|
|
},
|
|
"epoch_mean": 15,
|
|
"test_metrics": {
|
|
"size": {
|
|
"n_samples": 83,
|
|
"mse": 0.3297365176991437,
|
|
"rmse": 0.5742268869524865,
|
|
"mae": 0.3261373416486993,
|
|
"r2": -0.011808662805825154
|
|
},
|
|
"delivery": {
|
|
"n_samples": 59,
|
|
"mse": 0.7657906728091051,
|
|
"rmse": 0.8750946650557899,
|
|
"mae": 0.5747662091753998,
|
|
"r2": 0.2555604817459669
|
|
},
|
|
"pdi": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.6547619047619048,
|
|
"precision": 0.34004329004329004,
|
|
"recall": 0.5307219662058371,
|
|
"f1": 0.3543192918192918
|
|
},
|
|
"ee": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.6547619047619048,
|
|
"precision": 0.5846866096866097,
|
|
"recall": 0.6315736551030668,
|
|
"f1": 0.5823533191954245
|
|
},
|
|
"toxic": {
|
|
"n_samples": 60,
|
|
"accuracy": 0.9166666666666666,
|
|
"precision": 0.6875,
|
|
"recall": 0.956140350877193,
|
|
"f1": 0.749791492910759
|
|
},
|
|
"biodist": {
|
|
"n_samples": 60,
|
|
"kl_divergence": 0.2686512579596156,
|
|
"js_divergence": 0.06891811070906341
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"fold": 4,
|
|
"best_params": {
|
|
"dropout": 0.445253277169985,
|
|
"lr": 0.0002457483875420317,
|
|
"weight_decay": 0.00014672230276933923,
|
|
"backbone_lr_ratio": 0.2588275309142304,
|
|
"moe_n_experts": 2,
|
|
"moe_top_k": 2,
|
|
"moe_expert_hidden_mult": 2,
|
|
"d_model": 256,
|
|
"num_heads": 8,
|
|
"n_attn_layers": 4,
|
|
"fusion_strategy": "attention",
|
|
"head_hidden_dim": 128,
|
|
"set_transformer_block": "sab"
|
|
},
|
|
"epoch_mean": 25,
|
|
"test_metrics": {
|
|
"size": {
|
|
"n_samples": 84,
|
|
"mse": 0.29915782982104877,
|
|
"rmse": 0.5469532245275173,
|
|
"mae": 0.44856263058526175,
|
|
"r2": -1.1717654113653504
|
|
},
|
|
"delivery": {
|
|
"n_samples": 58,
|
|
"mse": 0.9258910073532772,
|
|
"rmse": 0.9622323042557225,
|
|
"mae": 0.6478613211676992,
|
|
"r2": 0.10467868059252439
|
|
},
|
|
"pdi": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.5833333333333334,
|
|
"precision": 0.328125,
|
|
"recall": 0.5144009216589862,
|
|
"f1": 0.3488413547237077
|
|
},
|
|
"ee": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.6785714285714286,
|
|
"precision": 0.6225589225589225,
|
|
"recall": 0.6371794871794871,
|
|
"f1": 0.6226530205687535
|
|
},
|
|
"toxic": {
|
|
"n_samples": 59,
|
|
"accuracy": 0.9661016949152542,
|
|
"precision": 0.8,
|
|
"recall": 0.9821428571428572,
|
|
"f1": 0.8659090909090909
|
|
},
|
|
"biodist": {
|
|
"n_samples": 58,
|
|
"kl_divergence": 0.36246349925173466,
|
|
"js_divergence": 0.09345889429140254
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"fold": 0,
|
|
"best_params": {
|
|
"dropout": 0.11659837578806788,
|
|
"lr": 0.00026604612647882186,
|
|
"weight_decay": 0.00021355962098733124,
|
|
"backbone_lr_ratio": 0.29733721720121176,
|
|
"moe_n_experts": 2,
|
|
"moe_top_k": 2,
|
|
"moe_expert_hidden_mult": 2,
|
|
"d_model": 256,
|
|
"num_heads": 8,
|
|
"n_attn_layers": 4,
|
|
"fusion_strategy": "attention",
|
|
"head_hidden_dim": 128,
|
|
"set_transformer_block": "sab"
|
|
},
|
|
"epoch_mean": 18,
|
|
"test_metrics": {
|
|
"size": {
|
|
"n_samples": 83,
|
|
"mse": 0.25539435269510624,
|
|
"rmse": 0.5053655634242467,
|
|
"mae": 0.2715319524328393,
|
|
"r2": 0.15199911066123628
|
|
},
|
|
"delivery": {
|
|
"n_samples": 58,
|
|
"mse": 0.5401788750549795,
|
|
"rmse": 0.7349686218165913,
|
|
"mae": 0.5700714552839254,
|
|
"r2": 0.3134000532957635
|
|
},
|
|
"pdi": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.5238095238095238,
|
|
"precision": 0.34578392621870885,
|
|
"recall": 0.53551912568306,
|
|
"f1": 0.34206626442256804
|
|
},
|
|
"ee": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.75,
|
|
"precision": 0.7166666666666667,
|
|
"recall": 0.7785714285714285,
|
|
"f1": 0.7271511627906978
|
|
},
|
|
"toxic": {
|
|
"n_samples": 58,
|
|
"accuracy": 0.9655172413793104,
|
|
"precision": 0.75,
|
|
"recall": 0.9821428571428572,
|
|
"f1": 0.8242424242424242
|
|
},
|
|
"biodist": {
|
|
"n_samples": 58,
|
|
"kl_divergence": 0.3182339907718673,
|
|
"js_divergence": 0.07309337216987152
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"fold": 1,
|
|
"best_params": {
|
|
"dropout": 0.18614946459458093,
|
|
"lr": 0.00015757063904422037,
|
|
"weight_decay": 0.000955120409040689,
|
|
"backbone_lr_ratio": 0.9059364477738808,
|
|
"moe_n_experts": 4,
|
|
"moe_top_k": 2,
|
|
"moe_expert_hidden_mult": 1,
|
|
"d_model": 256,
|
|
"num_heads": 8,
|
|
"n_attn_layers": 4,
|
|
"fusion_strategy": "attention",
|
|
"head_hidden_dim": 128,
|
|
"set_transformer_block": "sab"
|
|
},
|
|
"epoch_mean": 16,
|
|
"test_metrics": {
|
|
"size": {
|
|
"n_samples": 84,
|
|
"mse": 0.09045275471905934,
|
|
"rmse": 0.30075364456488196,
|
|
"mae": 0.23161025274367558,
|
|
"r2": -0.4984988024563053
|
|
},
|
|
"delivery": {
|
|
"n_samples": 61,
|
|
"mse": 1.0258191430256933,
|
|
"rmse": 1.0128273016786689,
|
|
"mae": 0.6263890392528694,
|
|
"r2": 0.2235288589357498
|
|
},
|
|
"pdi": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.5238095238095238,
|
|
"precision": 0.27182539682539686,
|
|
"recall": 0.25396825396825395,
|
|
"f1": 0.24962406015037591
|
|
},
|
|
"ee": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.5595238095238095,
|
|
"precision": 0.506110257723161,
|
|
"recall": 0.5057561486132914,
|
|
"f1": 0.49192340762639003
|
|
},
|
|
"toxic": {
|
|
"n_samples": 61,
|
|
"accuracy": 0.9508196721311475,
|
|
"precision": 0.75,
|
|
"recall": 0.9741379310344828,
|
|
"f1": 0.8200589970501475
|
|
},
|
|
"biodist": {
|
|
"n_samples": 61,
|
|
"kl_divergence": 0.46752297086632433,
|
|
"js_divergence": 0.11978345787287335
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"fold": 2,
|
|
"best_params": {
|
|
"dropout": 0.27066334445695367,
|
|
"lr": 0.00024715599869840884,
|
|
"weight_decay": 1.0172836375002765e-05,
|
|
"backbone_lr_ratio": 0.5054041560165133,
|
|
"moe_n_experts": 4,
|
|
"moe_top_k": 1,
|
|
"moe_expert_hidden_mult": 2,
|
|
"d_model": 256,
|
|
"num_heads": 8,
|
|
"n_attn_layers": 4,
|
|
"fusion_strategy": "attention",
|
|
"head_hidden_dim": 128,
|
|
"set_transformer_block": "sab"
|
|
},
|
|
"epoch_mean": 14,
|
|
"test_metrics": {
|
|
"size": {
|
|
"n_samples": 84,
|
|
"mse": 0.18305400438734884,
|
|
"rmse": 0.42784810901457637,
|
|
"mae": 0.33086523839405607,
|
|
"r2": -0.961365353255677
|
|
},
|
|
"delivery": {
|
|
"n_samples": 60,
|
|
"mse": 0.6782912801570451,
|
|
"rmse": 0.8235844098555078,
|
|
"mae": 0.5519710779888556,
|
|
"r2": 0.09707674202024197
|
|
},
|
|
"pdi": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.5476190476190477,
|
|
"precision": 0.4318260568260568,
|
|
"recall": 0.7012288786482334,
|
|
"f1": 0.4298268907350504
|
|
},
|
|
"ee": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.6547619047619048,
|
|
"precision": 0.5916666666666667,
|
|
"recall": 0.6414565826330533,
|
|
"f1": 0.6008547008547008
|
|
},
|
|
"toxic": {
|
|
"n_samples": 61,
|
|
"accuracy": 0.9344262295081968,
|
|
"precision": 0.7142857142857143,
|
|
"recall": 0.9655172413793103,
|
|
"f1": 0.7821428571428571
|
|
},
|
|
"biodist": {
|
|
"n_samples": 60,
|
|
"kl_divergence": 0.2848294483055189,
|
|
"js_divergence": 0.07664507104164577
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"fold": 3,
|
|
"best_params": {
|
|
"dropout": 0.2778201867797741,
|
|
"lr": 0.0004781288818954282,
|
|
"weight_decay": 0.0016987970387862138,
|
|
"backbone_lr_ratio": 0.9458696393367624,
|
|
"moe_n_experts": 4,
|
|
"moe_top_k": 2,
|
|
"moe_expert_hidden_mult": 2,
|
|
"d_model": 256,
|
|
"num_heads": 8,
|
|
"n_attn_layers": 4,
|
|
"fusion_strategy": "attention",
|
|
"head_hidden_dim": 128,
|
|
"set_transformer_block": "sab"
|
|
},
|
|
"epoch_mean": 15,
|
|
"test_metrics": {
|
|
"size": {
|
|
"n_samples": 83,
|
|
"mse": 0.3129556437133723,
|
|
"rmse": 0.5594243860553205,
|
|
"mae": 0.3127409808606987,
|
|
"r2": 0.039684067774135356
|
|
},
|
|
"delivery": {
|
|
"n_samples": 59,
|
|
"mse": 0.7857185201725363,
|
|
"rmse": 0.8864076489812891,
|
|
"mae": 0.6024630838626269,
|
|
"r2": 0.23618824646311343
|
|
},
|
|
"pdi": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.7023809523809523,
|
|
"precision": 0.38567649281934996,
|
|
"recall": 0.5862135176651306,
|
|
"f1": 0.40061936936936937
|
|
},
|
|
"ee": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.6428571428571429,
|
|
"precision": 0.5957671957671957,
|
|
"recall": 0.6363247863247863,
|
|
"f1": 0.5951323580633926
|
|
},
|
|
"toxic": {
|
|
"n_samples": 60,
|
|
"accuracy": 0.9333333333333333,
|
|
"precision": 0.7142857142857143,
|
|
"recall": 0.9649122807017544,
|
|
"f1": 0.7818181818181817
|
|
},
|
|
"biodist": {
|
|
"n_samples": 60,
|
|
"kl_divergence": 0.37168209467721186,
|
|
"js_divergence": 0.09777546387972921
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"fold": 4,
|
|
"best_params": {
|
|
"dropout": 0.445253277169985,
|
|
"lr": 0.0002457483875420317,
|
|
"weight_decay": 0.00014672230276933923,
|
|
"backbone_lr_ratio": 0.2588275309142304,
|
|
"moe_n_experts": 2,
|
|
"moe_top_k": 2,
|
|
"moe_expert_hidden_mult": 2,
|
|
"d_model": 256,
|
|
"num_heads": 8,
|
|
"n_attn_layers": 4,
|
|
"fusion_strategy": "attention",
|
|
"head_hidden_dim": 128,
|
|
"set_transformer_block": "sab"
|
|
},
|
|
"epoch_mean": 25,
|
|
"test_metrics": {
|
|
"size": {
|
|
"n_samples": 84,
|
|
"mse": 0.2967457163832042,
|
|
"rmse": 0.544743716240219,
|
|
"mae": 0.461613933245341,
|
|
"r2": -1.1542544388605225
|
|
},
|
|
"delivery": {
|
|
"n_samples": 58,
|
|
"mse": 0.9213812458325495,
|
|
"rmse": 0.9598860587760141,
|
|
"mae": 0.6561634901525646,
|
|
"r2": 0.10903954553546502
|
|
},
|
|
"pdi": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.5952380952380952,
|
|
"precision": 0.4396781819568367,
|
|
"recall": 0.7017409114183307,
|
|
"f1": 0.46197533353496656
|
|
},
|
|
"ee": {
|
|
"n_samples": 84,
|
|
"accuracy": 0.6666666666666666,
|
|
"precision": 0.6104279510076611,
|
|
"recall": 0.6205128205128205,
|
|
"f1": 0.610010886870937
|
|
},
|
|
"toxic": {
|
|
"n_samples": 59,
|
|
"accuracy": 0.9661016949152542,
|
|
"precision": 0.8,
|
|
"recall": 0.9821428571428572,
|
|
"f1": 0.8659090909090909
|
|
},
|
|
"biodist": {
|
|
"n_samples": 58,
|
|
"kl_divergence": 0.43138180820562744,
|
|
"js_divergence": 0.11081996861265214
|
|
}
|
|
}
|
|
}
|
|
],
|
|
"summary_stats": {
|
|
"size": {
|
|
"mse_mean": 0.23034984029037256,
|
|
"mse_std": 0.09049026510650632,
|
|
"rmse_mean": 0.46822986556281965,
|
|
"rmse_std": 0.105406988788203,
|
|
"mae_mean": 0.32189987007849846,
|
|
"mae_std": 0.08014105320611228,
|
|
"r2_mean": -0.4728185006647179,
|
|
"r2_std": 0.5427422370821634
|
|
},
|
|
"delivery": {
|
|
"mse_mean": 0.8054494222584121,
|
|
"mse_std": 0.16018870426337528,
|
|
"rmse_mean": 0.8929394867547544,
|
|
"rmse_std": 0.09004718347937392,
|
|
"mae_mean": 0.6145264492845487,
|
|
"mae_std": 0.045144810651025415,
|
|
"r2_mean": 0.17842704120419808,
|
|
"r2_std": 0.06422601059093863
|
|
},
|
|
"pdi": {
|
|
"accuracy_mean": 0.5896825396825397,
|
|
"accuracy_std": 0.05682229307831347,
|
|
"precision_mean": 0.35469393906374463,
|
|
"precision_std": 0.055592769200553505,
|
|
"recall_mean": 0.522735863838455,
|
|
"recall_std": 0.15050332076418682,
|
|
"f1_mean": 0.3604048512722805,
|
|
"f1_std": 0.06427235489398077
|
|
},
|
|
"ee": {
|
|
"accuracy_mean": 0.6547619047619047,
|
|
"accuracy_std": 0.05481364015819684,
|
|
"precision_mean": 0.6019542483742392,
|
|
"precision_std": 0.06379027980318479,
|
|
"recall_mean": 0.6352859185632295,
|
|
"recall_std": 0.08160957226119422,
|
|
"f1_mean": 0.6021181253673275,
|
|
"f1_std": 0.06960400235208303
|
|
},
|
|
"toxic": {
|
|
"accuracy_mean": 0.9500779484296937,
|
|
"accuracy_std": 0.017755203225104002,
|
|
"precision_mean": 0.7600104427736006,
|
|
"precision_std": 0.07046117058026663,
|
|
"recall_mean": 0.9577096476824245,
|
|
"recall_std": 0.05611117888622105,
|
|
"f1_mean": 0.813010272994035,
|
|
"f1_std": 0.03449626596837632
|
|
},
|
|
"biodist": {
|
|
"kl_divergence_mean": 0.3520444477507348,
|
|
"kl_divergence_std": 0.08432681015154675,
|
|
"js_divergence_mean": 0.08975023050906776,
|
|
"js_divergence_std": 0.021944673828634763
|
|
}
|
|
}
|
|
} |