indonesian-roberta-base-posp-tagger-finetuned-indonlu-smsa

This model is a fine-tuned version of w11wo/indonesian-roberta-base-posp-tagger on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.6399
• Balanced Accuracy: 0.6249

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: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1000
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Balanced Accuracy
No log	1.0	250	1.0331	0.2308
1.1193	2.0	500	0.8059	0.4522
1.1193	3.0	750	0.8079	0.4752
0.6263	4.0	1000	0.8437	0.3973
0.6263	5.0	1250	0.9804	0.5115
0.3579	6.0	1500	1.1114	0.5958
0.3579	7.0	1750	1.3259	0.5683
0.2035	8.0	2000	1.3290	0.5328
0.2035	9.0	2250	1.4968	0.5712
0.1192	10.0	2500	1.4737	0.5299
0.1192	11.0	2750	1.5556	0.5513
0.0801	12.0	3000	1.6399	0.6249
0.0801	13.0	3250	1.5364	0.5677
0.049	14.0	3500	1.5462	0.5638
0.049	15.0	3750	1.5863	0.5682

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1