bert-base-indonesian-1.5G-multilabel-indonesian-hate-speech-modified

This model is a fine-tuned version of cahya/bert-base-indonesian-1.5G on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2368
• F1: 0.8123
• Roc Auc: 0.8828
• Accuracy: 0.7396

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.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: 15

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
0.2277	1.0	1317	0.1909	0.7771	0.8519	0.6659
0.1552	2.0	2634	0.1908	0.7910	0.8718	0.6735
0.0983	3.0	3951	0.2215	0.7898	0.8681	0.7183
0.0579	4.0	5268	0.2368	0.8123	0.8828	0.7396
0.0389	5.0	6585	0.2813	0.8039	0.8815	0.7327
0.0264	6.0	7902	0.3052	0.7992	0.8788	0.7396
0.0191	7.0	9219	0.3389	0.7976	0.8791	0.7365
0.0122	8.0	10536	0.3407	0.8041	0.8797	0.7373
0.0076	9.0	11853	0.3596	0.8014	0.8821	0.7373
0.007	10.0	13170	0.3688	0.8086	0.8861	0.7456
0.0049	11.0	14487	0.3850	0.8006	0.8743	0.7487
0.0039	12.0	15804	0.3863	0.8085	0.8850	0.7472
0.003	13.0	17121	0.3884	0.8080	0.8766	0.7555
0.0016	14.0	18438	0.3909	0.8102	0.8872	0.7509
0.0019	15.0	19755	0.3929	0.8083	0.8819	0.7509

Framework versions

• Transformers 4.51.3
• Pytorch 2.7.0+cu128
• Datasets 3.6.0
• Tokenizers 0.21.1