README.md

metadata

library_name: transformers
license: mit
base_model: cahya/bert-base-indonesian-1.5G
tags:
  - generated_from_trainer
metrics:
  - f1
  - accuracy
model-index:
  - name: bert-base-indonesian-1.5G-multilabel-indonesian-hate-speech-new
    results: []

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

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: 0.3641
• F1: 0.7802
• Roc Auc: 0.8639
• Accuracy: 0.7156

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
0.3106	1.0	659	0.2504	0.6779	0.7832	0.5978
0.2235	2.0	1318	0.2113	0.7466	0.8392	0.6441
0.1722	3.0	1977	0.2283	0.7511	0.8493	0.6581
0.097	4.0	2636	0.2421	0.7626	0.8490	0.6874
0.0643	5.0	3295	0.2727	0.7584	0.8417	0.6938
0.0572	6.0	3954	0.2817	0.7662	0.8662	0.6737
0.0304	7.0	4613	0.3075	0.7606	0.8475	0.6879
0.021	8.0	5272	0.3195	0.7697	0.8626	0.6932
0.0157	9.0	5931	0.3347	0.7663	0.8477	0.7052
0.0095	10.0	6590	0.3353	0.7759	0.8598	0.7118
0.0086	11.0	7249	0.3467	0.7768	0.8590	0.7136
0.0063	12.0	7908	0.3503	0.7795	0.8644	0.7128
0.0046	13.0	8567	0.3577	0.7797	0.8613	0.7153
0.0037	14.0	9226	0.3622	0.7801	0.8674	0.7115
0.0046	15.0	9885	0.3641	0.7802	0.8639	0.7156

Framework versions

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