RoBERTa-Base-SE2025T11A-sun-v20250110155704

This model is a fine-tuned version of w11wo/sundanese-roberta-base-emotion-classifier on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3025
• F1 Macro: 0.4497
• F1 Micro: 0.5706
• F1 Weighted: 0.5077
• F1 Samples: 0.4977
• F1 Label Marah: 0.5546
• F1 Label Jijik: 0.1
• F1 Label Takut: 0.5263
• F1 Label Senang: 0.8163
• F1 Label Sedih: 0.768
• F1 Label Terkejut: 0.3830
• F1 Label Biasa: 0.0

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: 2e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss	F1 Macro	F1 Micro	F1 Weighted	F1 Samples	F1 Label Marah	F1 Label Jijik	F1 Label Takut	F1 Label Senang	F1 Label Sedih	F1 Label Terkejut	F1 Label Biasa
0.4916	0.1133	100	0.4178	0.2021	0.3541	0.2559	0.2595	0.0	0.0	0.2	0.74	0.4746	0.0	0.0
0.4255	0.2265	200	0.3776	0.2103	0.3860	0.2716	0.2849	0.1429	0.0	0.0364	0.7982	0.4946	0.0	0.0
0.3773	0.3398	300	0.3585	0.2870	0.4140	0.3408	0.3114	0.2821	0.0	0.3636	0.7732	0.4835	0.1067	0.0
0.3915	0.4530	400	0.3399	0.3393	0.4488	0.3897	0.3411	0.24	0.0952	0.4658	0.7403	0.6542	0.1795	0.0
0.3673	0.5663	500	0.3209	0.4097	0.5318	0.4661	0.4475	0.5	0.0	0.5385	0.7919	0.6964	0.3409	0.0
0.3436	0.6795	600	0.3169	0.4393	0.5431	0.4927	0.4589	0.5	0.1270	0.5455	0.7624	0.7402	0.4	0.0
0.3203	0.7928	700	0.3085	0.4193	0.5430	0.4766	0.4577	0.4615	0.0345	0.5641	0.8079	0.6897	0.3778	0.0
0.3507	0.9060	800	0.3025	0.4497	0.5706	0.5077	0.4977	0.5546	0.1	0.5263	0.8163	0.768	0.3830	0.0

Framework versions

• Transformers 4.48.0
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0