hubert-base-superb-er-3kfoldfull40-finetuned-bmd-20250824_112936-LOSO-section-out1

This model is a fine-tuned version of superb/hubert-base-superb-er on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.9806
• Accuracy: 0.5417
• F1: 0.4444

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e-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 results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	1.0	3	1.0946	0.3333	0.1667
No log	2.0	6	1.0941	0.3333	0.1667
No log	3.0	9	1.0928	0.3333	0.1839
1.0958	4.0	12	1.0911	0.375	0.2778
1.0958	5.0	15	1.0893	0.4583	0.3626
1.0958	6.0	18	1.0877	0.4167	0.3375
1.0877	7.0	21	1.0861	0.5417	0.4533
1.0877	8.0	24	1.0840	0.5	0.4176
1.0877	9.0	27	1.0814	0.5	0.4176
1.0764	10.0	30	1.0781	0.5	0.4176
1.0764	11.0	33	1.0744	0.5	0.4176
1.0764	12.0	36	1.0694	0.5417	0.4539
1.0764	13.0	39	1.0638	0.5417	0.4539
1.0655	14.0	42	1.0571	0.5417	0.4539
1.0655	15.0	45	1.0501	0.5417	0.4539
1.0655	16.0	48	1.0426	0.5	0.4127
1.0437	17.0	51	1.0357	0.5	0.4127
1.0437	18.0	54	1.0292	0.5	0.4127
1.0437	19.0	57	1.0221	0.4583	0.3780
1.0197	20.0	60	1.0151	0.5	0.4176
1.0197	21.0	63	1.0088	0.5417	0.4533
1.0197	22.0	66	1.0039	0.5417	0.4533
1.0197	23.0	69	0.9989	0.5417	0.4533
0.9988	24.0	72	0.9927	0.5417	0.4533
0.9988	25.0	75	0.9884	0.5417	0.4533
0.9988	26.0	78	0.9851	0.5417	0.4533
0.9709	27.0	81	0.9839	0.5417	0.4533
0.9709	28.0	84	0.9816	0.5417	0.4444
0.9709	29.0	87	0.9806	0.5417	0.4444
0.9614	30.0	90	0.9806	0.5417	0.4444

Framework versions

• Transformers 4.55.2
• Pytorch 2.8.0+cu126
• Datasets 3.6.0
• Tokenizers 0.21.4