{ "fold_results": [ { "fold": 0, "best_params": { "dropout": 0.19323529834582587, "lr": 0.000980484886752915, "weight_decay": 0.00014367509113739864, "backbone_lr_ratio": 0.13093310334961422, "moe_n_experts": 4, "moe_top_k": 2, "moe_expert_hidden_mult": 2, "llm_lora_r": 8, "d_model": 256, "num_heads": 8, "n_attn_layers": 4, "fusion_strategy": "attention", "head_hidden_dim": 128, "set_transformer_block": "sab" }, "epoch_mean": 11, "test_metrics": { "size": { "n_samples": 83, "mse": 1.4858819540561683, "rmse": 1.2189675771144073, "mae": 0.5309677058494234, "r2": 0.0906036692787614 }, "delivery": { "n_samples": 58, "mse": 0.4578579988449137, "rmse": 0.6766520515338099, "mae": 0.44371995239936074, "r2": 0.41803485452290234 }, "pdi": { "n_samples": 84, "accuracy": 0.7857142857142857, "precision": 0.7322033898305085, "recall": 0.7441197434069851, "f1": 0.7375 }, "ee": { "n_samples": 84, "accuracy": 0.7380952380952381, "precision": 0.7146208708708709, "recall": 0.7547619047619047, "f1": 0.7160330091364574 }, "toxic": { "n_samples": 58, "accuracy": 0.9827586206896551, "precision": 0.9912280701754386, "recall": 0.75, "f1": 0.8289085545722714 }, "biodist": { "n_samples": 58, "kl_divergence": 0.2124931399005794, "js_divergence": 0.0436771104854127 } } }, { "fold": 1, "best_params": { "dropout": 0.19323529834582587, "lr": 0.000980484886752915, "weight_decay": 0.00014367509113739864, "backbone_lr_ratio": 0.13093310334961422, "moe_n_experts": 4, "moe_top_k": 2, "moe_expert_hidden_mult": 2, "llm_lora_r": 8, "d_model": 256, "num_heads": 8, "n_attn_layers": 4, "fusion_strategy": "attention", "head_hidden_dim": 128, "set_transformer_block": "sab" }, "epoch_mean": 11, "test_metrics": { "size": { "n_samples": 84, "mse": 0.3521366695033162, "rmse": 0.593411045990312, "mae": 0.417994069290303, "r2": -0.07530019609216088 }, "delivery": { "n_samples": 61, "mse": 0.8615718355352413, "rmse": 0.9282089395902419, "mae": 0.54733644703739, "r2": 0.34785223029308054 }, "pdi": { "n_samples": 84, "accuracy": 0.6309523809523809, "precision": 0.5417771883289124, "recall": 0.5476190476190477, "f1": 0.5421135924037278 }, "ee": { "n_samples": 84, "accuracy": 0.5952380952380952, "precision": 0.531971296677179, "recall": 0.5323723895152467, "f1": 0.5202020202020202 }, "toxic": { "n_samples": 61, "accuracy": 0.9836065573770492, "precision": 0.9915254237288136, "recall": 0.8333333333333333, "f1": 0.8957264957264958 }, "biodist": { "n_samples": 61, "kl_divergence": 0.12922332899139097, "js_divergence": 0.030354793549857463 } } }, { "fold": 2, "best_params": { "dropout": 0.19323529834582587, "lr": 0.000980484886752915, "weight_decay": 0.00014367509113739864, "backbone_lr_ratio": 0.13093310334961422, "moe_n_experts": 4, "moe_top_k": 2, "moe_expert_hidden_mult": 2, "llm_lora_r": 8, "d_model": 256, "num_heads": 8, "n_attn_layers": 4, "fusion_strategy": "attention", "head_hidden_dim": 128, "set_transformer_block": "sab" }, "epoch_mean": 11, "test_metrics": { "size": { "n_samples": 84, "mse": 0.4170957202870868, "rmse": 0.6458294823613171, "mae": 0.4985973424876907, "r2": 0.1762434046684933 }, "delivery": { "n_samples": 60, "mse": 0.5854544655957109, "rmse": 0.7651499628149445, "mae": 0.49963968732239056, "r2": 0.22065863304023947 }, "pdi": { "n_samples": 84, "accuracy": 0.7380952380952381, "precision": 0.6851851851851851, "recall": 0.7199413489736071, "f1": 0.693633952254642 }, "ee": { "n_samples": 84, "accuracy": 0.6666666666666666, "precision": 0.613035584604212, "recall": 0.6666666666666666, "f1": 0.627436739505705 }, "toxic": { "n_samples": 61, "accuracy": 1.0, "precision": 1.0, "recall": 1.0, "f1": 1.0 }, "biodist": { "n_samples": 60, "kl_divergence": 0.17218262970209255, "js_divergence": 0.039882423974733884 } } }, { "fold": 3, "best_params": { "dropout": 0.19323529834582587, "lr": 0.000980484886752915, "weight_decay": 0.00014367509113739864, "backbone_lr_ratio": 0.13093310334961422, "moe_n_experts": 4, "moe_top_k": 2, "moe_expert_hidden_mult": 2, "llm_lora_r": 8, "d_model": 256, "num_heads": 8, "n_attn_layers": 4, "fusion_strategy": "attention", "head_hidden_dim": 128, "set_transformer_block": "sab" }, "epoch_mean": 11, "test_metrics": { "size": { "n_samples": 83, "mse": 1.773779594197195, "rmse": 1.331833170557482, "mae": 0.5606547438887408, "r2": -0.003263258057283336 }, "delivery": { "n_samples": 59, "mse": 0.6480387215025862, "rmse": 0.8050085226272988, "mae": 0.5227914613403254, "r2": 0.37002936863191194 }, "pdi": { "n_samples": 84, "accuracy": 0.75, "precision": 0.6789023521026372, "recall": 0.684017595307918, "f1": 0.6813008130081302 }, "ee": { "n_samples": 84, "accuracy": 0.6428571428571429, "precision": 0.5700750584471516, "recall": 0.5891402714932127, "f1": 0.5736800630417652 }, "toxic": { "n_samples": 60, "accuracy": 0.9833333333333333, "precision": 0.875, "recall": 0.9912280701754386, "f1": 0.9241466498103665 }, "biodist": { "n_samples": 60, "kl_divergence": 0.15399347135454697, "js_divergence": 0.036819733414228055 } } }, { "fold": 4, "best_params": { "dropout": 0.19323529834582587, "lr": 0.000980484886752915, "weight_decay": 0.00014367509113739864, "backbone_lr_ratio": 0.13093310334961422, "moe_n_experts": 4, "moe_top_k": 2, "moe_expert_hidden_mult": 2, "llm_lora_r": 8, "d_model": 256, "num_heads": 8, "n_attn_layers": 4, "fusion_strategy": "attention", "head_hidden_dim": 128, "set_transformer_block": "sab" }, "epoch_mean": 11, "test_metrics": { "size": { "n_samples": 84, "mse": 0.6805742520206829, "rmse": 0.8249692430755724, "mae": 0.4822031688111435, "r2": 0.08930711264607172 }, "delivery": { "n_samples": 58, "mse": 0.7669285932160084, "rmse": 0.8757445935979328, "mae": 0.567918715805843, "r2": 0.2583927108955246 }, "pdi": { "n_samples": 84, "accuracy": 0.7619047619047619, "precision": 0.7002652519893899, "recall": 0.7214076246334311, "f1": 0.7083333333333334 }, "ee": { "n_samples": 84, "accuracy": 0.7023809523809523, "precision": 0.6327380952380953, "recall": 0.6816239316239318, "f1": 0.6431159420289855 }, "toxic": { "n_samples": 59, "accuracy": 1.0, "precision": 1.0, "recall": 1.0, "f1": 1.0 }, "biodist": { "n_samples": 58, "kl_divergence": 0.16534984510664474, "js_divergence": 0.041883856181021084 } } } ], "summary_stats": { "size": { "mse_mean": 0.9418936380128897, "mse_std": 0.5795638860278585, "rmse_mean": 0.9230021038198182, "rmse_std": 0.29993458346292695, "mae_mean": 0.4980834060654603, "mae_std": 0.048297613647592774, "r2_mean": 0.05551814648877644, "r2_std": 0.0866209771152643 }, "delivery": { "mse_mean": 0.6639703229388921, "mse_std": 0.14039060447259435, "rmse_mean": 0.8101528140328454, "rmse_std": 0.08730830918963939, "mae_mean": 0.5162812527810619, "mae_std": 0.04293092010991082, "r2_mean": 0.3229935594767318, "r2_std": 0.07281328920995907 }, "pdi": { "accuracy_mean": 0.7333333333333333, "accuracy_std": 0.05355819942477139, "precision_mean": 0.6676666734873267, "precision_std": 0.06558804201410597, "recall_mean": 0.6834210719881977, "recall_std": 0.07057419243292298, "f1_mean": 0.6725763381999668, "f1_std": 0.06787298751865974 }, "ee": { "accuracy_mean": 0.669047619047619, "accuracy_std": 0.04902681019517619, "precision_mean": 0.6124881811675018, "precision_std": 0.061843813988713685, "recall_mean": 0.6449130328121926, "recall_std": 0.07702425935365928, "f1_mean": 0.6160935547829867, "f1_std": 0.06609215153794168 }, "toxic": { "accuracy_mean": 0.9899397022800075, "accuracy_std": 0.008218758400859814, "precision_mean": 0.9715506987808504, "precision_std": 0.04842922980294785, "recall_mean": 0.9149122807017542, "recall_std": 0.10407221017518521, "f1_mean": 0.9297563400218267, "f1_std": 0.06515845039804279 }, "biodist": { "kl_divergence_mean": 0.16664848301105092, "kl_divergence_std": 0.027183955851732975, "js_divergence_mean": 0.03852358352105063, "js_divergence_std": 0.004676087139485391 } } }