nerugm-unipelt

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: 0.1898
• Precision: 0.8077
• Recall: 0.8861
• F1: 0.8451
• Accuracy: 0.9649

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
0.4277	1.0	528	0.1555	0.6979	0.8576	0.7696	0.9475
0.1337	2.0	1056	0.1242	0.7596	0.8704	0.8113	0.9584
0.103	3.0	1584	0.1320	0.7606	0.8768	0.8146	0.9593
0.085	4.0	2112	0.1275	0.7738	0.8745	0.8211	0.9601
0.0708	5.0	2640	0.1263	0.7849	0.8797	0.8296	0.9636
0.0613	6.0	3168	0.1332	0.8060	0.8884	0.8452	0.9646
0.0529	7.0	3696	0.1474	0.7720	0.8815	0.8231	0.9595
0.0462	8.0	4224	0.1607	0.7690	0.8861	0.8234	0.9584
0.0409	9.0	4752	0.1463	0.7881	0.8687	0.8264	0.9643
0.0351	10.0	5280	0.1562	0.8019	0.8704	0.8348	0.9631
0.0328	11.0	5808	0.1607	0.7931	0.8908	0.8391	0.9640
0.0286	12.0	6336	0.1701	0.8077	0.8884	0.8462	0.9646
0.0262	13.0	6864	0.1667	0.8	0.8855	0.8406	0.9654
0.0241	14.0	7392	0.1702	0.8149	0.8954	0.8533	0.9659
0.0218	15.0	7920	0.1833	0.8022	0.8861	0.8421	0.9644
0.0193	16.0	8448	0.1767	0.8163	0.8884	0.8509	0.9667
0.0199	17.0	8976	0.1859	0.8009	0.8884	0.8424	0.9638
0.0175	18.0	9504	0.1842	0.8133	0.8861	0.8482	0.9665
0.0165	19.0	10032	0.1883	0.8099	0.8838	0.8452	0.9649
0.0164	20.0	10560	0.1898	0.8077	0.8861	0.8451	0.9649

Framework versions

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