indonesia-election-topic-classification-undersampling-double

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

• Loss: 2.2574
• Balanced Accuracy: 0.5978

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 14

Training results

Training Loss	Epoch	Step	Validation Loss	Balanced Accuracy
No log	1.0	180	1.0928	0.4716
No log	2.0	360	1.1167	0.4998
0.997	3.0	540	1.2231	0.4872
0.997	4.0	720	1.4308	0.5272
0.997	5.0	900	1.7431	0.5498
0.291	6.0	1080	1.9512	0.5124
0.291	7.0	1260	2.0338	0.5513
0.291	8.0	1440	2.1909	0.5427
0.1077	9.0	1620	2.2574	0.5978
0.1077	10.0	1800	2.3393	0.5698
0.1077	11.0	1980	2.3479	0.5312
0.0485	12.0	2160	2.3941	0.5795
0.0485	13.0	2340	2.4026	0.5896
0.0178	14.0	2520	2.4172	0.5584

Framework versions

• Transformers 4.41.2
• Pytorch 2.1.2
• Datasets 2.19.2
• Tokenizers 0.19.1