Unified Model Summary MD-only

Kolom: Accuracy, Macro F1, Weighted F1, Best Val Acc. Identifier: Transfer/CNN = Exp ID + Model ID; MLP = Run/Tag. File bernama 'best' tidak ditampilkan pada tabel.

🏆 Best Model Overall

Model: transferlearning | Exp ID: exp_tl_001 • Model ID: MNv3l

Accuracy: 99.64% Macro F1: 99.64% Weighted F1: 99.64% Best Val Acc: 99.33%

Open detail • Open MD

CNN

model	Exp ID	Model ID	accuracy	macro_f1	weighted_f1	best_val_acc	detail	md
Cnn	exp_004	D	98.56%	98.56%	98.56%	98.92%	detail	md
Cnn	exp_002	B	81.44%	81.73%	81.73%	82.00%	detail	md
Cnn	exp_005	E	80.21%	80.01%	80.01%	79.44%	detail	md
Cnn	exp_001	A	78.15%	78.36%	78.36%	79.28%	detail	md

Transfer Learning

model	Exp ID	Model ID	accuracy	macro_f1	weighted_f1	best_val_acc	detail	md
transferlearning	exp_tl_001	MNv3l	99.64%	99.64%	99.64%	99.33%	detail	md
transferlearning	exp_tl_004	ResNet50	99.59%	99.59%	99.59%	99.28%	detail	md
transferlearning	exp_tl_002	EffB0	99.49%	99.49%	99.49%	99.38%	detail	md
transferlearning	exp_tl_003	MNv2	99.33%	99.33%	99.33%	99.38%	detail	md
transferlearning	exp_tl_005	Xception	99.28%	99.28%	99.28%	99.49%	detail	md
transferlearning	exp_tl_006	InceptionV3	98.97%	98.97%	98.97%	98.92%	detail	md
transferlearning	exp_tl_008	VGG16	98.77%	98.77%	98.77%	98.82%	detail	md
transferlearning	exp_tl_007	MNv3s	98.67%	98.67%	98.67%	98.67%	detail	md

MLP — Experiments

model	Run/Tag	accuracy	macro_f1	weighted_f1	best_val_acc	detail	md
Mlp	user_reference_adam_100e_b128	96.90%	96.75%	96.90%	97.86%	detail	md
Mlp	high_lr_fast	96.71%	96.53%	96.71%	97.90%	detail	md
Mlp	balanced_long	96.42%	96.30%	96.41%	97.77%	detail	md
Mlp	swish_deep	96.38%	96.23%	96.36%	97.86%	detail	md
Mlp	deep_3layers_768_384_192	96.28%	96.17%	96.26%	97.68%	detail	md
Mlp	big_batch_512	96.23%	96.10%	96.22%	97.46%	detail	md
Mlp	narrow_many_4layers	96.23%	96.08%	96.23%	97.64%	detail	md
Mlp	baseline_large_batch	96.23%	96.07%	96.22%	97.73%	detail	md
Mlp	elu_3layers	96.18%	96.03%	96.17%	97.59%	detail	md
Mlp	rmsprop_mid	96.14%	96.01%	96.12%	97.68%	detail	md
Mlp	lightweight_small	95.99%	95.85%	95.99%	97.38%	detail	md
Mlp	wide_fast_b512	95.80%	95.69%	95.78%	97.38%	detail	md
Mlp	sgd_nesterov_long	95.37%	95.24%	95.34%	97.33%	detail	md
Mlp	swish_low_lr	95.42%	95.20%	95.41%	97.38%	detail	md
Mlp	no_reduce_lr	95.37%	95.17%	95.34%	97.46%	detail	md
Mlp	no_smote_adam	95.33%	95.12%	95.32%	96.30%	detail	md
Mlp	strong_dropout	95.23%	95.10%	95.18%	97.33%	detail	md
Mlp	adamw_swish_bigbatch	94.85%	94.68%	94.80%	97.29%	detail	md
Mlp	sgd_conservative	94.56%	94.29%	94.51%	94.67%	detail	md
Mlp	rmsprop_bigbatch	94.33%	94.04%	94.27%	94.50%	detail	md

MLP — Hyperparameters grid_base_ref_*

model	Run/Tag	accuracy	macro_f1	weighted_f1	best_val_acc	detail	md
Mlp	grid_base_ref_i033	96.80%	96.62%	96.80%	97.86%	detail	md
Mlp	grid_base_ref_i035	96.76%	96.61%	96.76%	97.99%	detail	md
Mlp	grid_base_ref_i027	96.61%	96.49%	96.61%	97.86%	detail	md
Mlp	grid_base_ref_i029	96.57%	96.43%	96.56%	97.73%	detail	md
Mlp	grid_base_ref_i021	96.47%	96.30%	96.48%	97.77%	detail	md
Mlp	grid_base_ref_i013	96.47%	96.27%	96.48%	97.68%	detail	md
Mlp	grid_base_ref_i019	96.42%	96.27%	96.42%	97.81%	detail	md
Mlp	grid_base_ref_i001	96.42%	96.26%	96.41%	97.73%	detail	md
Mlp	grid_base_ref_i009	96.38%	96.24%	96.36%	97.64%	detail	md
Mlp	grid_base_ref_i023	96.38%	96.20%	96.38%	97.77%	detail	md