scb-finetune-bert

This model is a fine-tuned version of google-bert/bert-base-multilingual-cased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.3597
• Accuracy: 0.3558
• Precision: 0.2910
• Recall: 0.3558
• F1: 0.3201

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: 16
• eval_batch_size: 16
• 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: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
No log	1.0	15	1.0511	0.375	0.1406	0.375	0.2045
No log	2.0	30	1.0649	0.375	0.1406	0.375	0.2045
No log	3.0	45	1.0597	0.3558	0.2953	0.3558	0.2957
No log	4.0	60	1.0902	0.3269	0.2611	0.3269	0.2650
No log	5.0	75	1.0790	0.4135	0.3320	0.4135	0.3657
No log	6.0	90	1.1469	0.3846	0.3139	0.3846	0.3457
No log	7.0	105	1.2955	0.3462	0.2866	0.3462	0.3015
No log	8.0	120	1.3283	0.3269	0.2717	0.3269	0.2933
No log	9.0	135	1.3397	0.375	0.3036	0.375	0.3351
No log	10.0	150	1.3597	0.3558	0.2910	0.3558	0.3201

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.5.1
• Tokenizers 0.21.1