nerugm-lora-r2a2d0.15

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: 0.1346
• Precision: 0.7342
• Recall: 0.8652
• F1: 0.7943
• Accuracy: 0.9555

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
0.79	1.0	528	0.4638	0.3302	0.0813	0.1305	0.8595
0.3919	2.0	1056	0.2519	0.5954	0.6729	0.6318	0.9275
0.2386	3.0	1584	0.1927	0.6540	0.7908	0.7159	0.9382
0.193	4.0	2112	0.1677	0.6826	0.8234	0.7464	0.9448
0.1712	5.0	2640	0.1594	0.6959	0.8443	0.7629	0.9476
0.1596	6.0	3168	0.1544	0.7082	0.8559	0.7751	0.9498
0.1524	7.0	3696	0.1519	0.7012	0.8605	0.7728	0.9506
0.1452	8.0	4224	0.1461	0.7203	0.8605	0.7842	0.9522
0.1397	9.0	4752	0.1432	0.7263	0.8559	0.7858	0.9535
0.1369	10.0	5280	0.1394	0.7258	0.8536	0.7845	0.9539
0.1336	11.0	5808	0.1375	0.7321	0.8512	0.7872	0.9543
0.1305	12.0	6336	0.1375	0.7345	0.8536	0.7896	0.9547
0.1281	13.0	6864	0.1351	0.7330	0.8536	0.7887	0.9547
0.1252	14.0	7392	0.1360	0.7342	0.8652	0.7943	0.9553
0.124	15.0	7920	0.1364	0.7292	0.8559	0.7875	0.9541
0.1234	16.0	8448	0.1351	0.7260	0.8605	0.7876	0.9549
0.1224	17.0	8976	0.1357	0.7299	0.8652	0.7918	0.9549
0.1208	18.0	9504	0.1360	0.7333	0.8675	0.7948	0.9553
0.1201	19.0	10032	0.1350	0.7347	0.8675	0.7956	0.9555
0.1205	20.0	10560	0.1346	0.7342	0.8652	0.7943	0.9555

Framework versions

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