nerugm-lora-rad

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

• Loss: 0.1778
• Precision: 0.7975
• Recall: 0.8698
• F1: 0.8321
• Accuracy: 0.9608

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: 5e-05
• train_batch_size: 16
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20.0

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.4434	1.0	528	0.1630	0.6799	0.8629	0.7606	0.9456
0.1462	2.0	1056	0.1294	0.7481	0.8768	0.8074	0.9567
0.1183	3.0	1584	0.1378	0.7521	0.8815	0.8117	0.9569
0.1012	4.0	2112	0.1359	0.7720	0.8815	0.8231	0.9597
0.0884	5.0	2640	0.1266	0.7930	0.8815	0.8349	0.9622
0.0793	6.0	3168	0.1409	0.8031	0.8815	0.8404	0.9610
0.072	7.0	3696	0.1546	0.7704	0.8815	0.8222	0.9589
0.067	8.0	4224	0.1433	0.7980	0.8722	0.8334	0.9608
0.0607	9.0	4752	0.1468	0.7864	0.8815	0.8312	0.9599
0.0562	10.0	5280	0.1497	0.7783	0.8815	0.8267	0.9612
0.0506	11.0	5808	0.1600	0.7938	0.8768	0.8332	0.9595
0.0483	12.0	6336	0.1596	0.7950	0.8745	0.8329	0.9608
0.0443	13.0	6864	0.1596	0.7786	0.8745	0.8238	0.9606
0.0421	14.0	7392	0.1650	0.7971	0.8768	0.8351	0.9612
0.0395	15.0	7920	0.1693	0.7908	0.8698	0.8284	0.9603
0.0375	16.0	8448	0.1725	0.7926	0.8791	0.8336	0.9595
0.0358	17.0	8976	0.1789	0.7975	0.8698	0.8321	0.9612
0.0339	18.0	9504	0.1782	0.7821	0.8675	0.8226	0.9601
0.0327	19.0	10032	0.1743	0.8010	0.8698	0.8340	0.9620
0.0327	20.0	10560	0.1778	0.7975	0.8698	0.8321	0.9608

Framework versions

• Transformers 4.39.3
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.15.2