learning_rate: 2e-05
train_batch_size: 8
eval_batch_size: 8
seed: 1968
gradient_accumulation_steps: 4
total_train_batch_size: 32
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 30
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	1.0	3	1.1069	0.3333	0.1667
No log	2.0	6	1.1057	0.3333	0.1667
No log	3.0	9	1.1031	0.375	0.2461
1.1006	4.0	12	1.1003	0.2917	0.2169
1.1006	5.0	15	1.0982	0.3333	0.2626
1.1006	6.0	18	1.0964	0.25	0.1778
1.0878	7.0	21	1.0944	0.3333	0.2775
1.0878	8.0	24	1.0911	0.375	0.3017
1.0878	9.0	27	1.0875	0.3333	0.2296
1.0708	10.0	30	1.0825	0.3333	0.2296
1.0708	11.0	33	1.0752	0.375	0.3017
1.0708	12.0	36	1.0615	0.375	0.3333
1.0708	13.0	39	1.0489	0.5417	0.5462
1.0439	14.0	42	1.0394	0.5417	0.5462
1.0439	15.0	45	1.0201	0.5417	0.5587
1.0439	16.0	48	1.0012	0.5417	0.5587
0.9911	17.0	51	0.9850	0.5833	0.5940
0.9911	18.0	54	0.9805	0.5417	0.5543
0.9911	19.0	57	0.9663	0.5833	0.5940
0.9186	20.0	60	0.9603	0.5833	0.5940
0.9186	21.0	63	0.9539	0.5417	0.5530
0.9186	22.0	66	0.9497	0.5417	0.5530
0.9186	23.0	69	0.9460	0.5417	0.5530
0.8535	24.0	72	0.9458	0.5417	0.5530
0.8535	25.0	75	0.9438	0.5417	0.5530
0.8535	26.0	78	0.9382	0.5417	0.5530
0.7935	27.0	81	0.9371	0.5417	0.5530
0.7935	28.0	84	0.9357	0.5417	0.5530
0.7935	29.0	87	0.9346	0.5417	0.5530
0.7683	30.0	90	0.9340	0.5417	0.5530

sarasarasara/hubert-base-superb-er-3kfoldfull30-V3-finetuned-bmd-20250824_123804-LOSO-section-out1

hubert-base-superb-er-3kfoldfull30-V3-finetuned-bmd-20250824_123804-LOSO-section-out1