CS221-xlm-roberta-base-ptbr-finetuned-finetuned-ptbr-tapt

This model is a fine-tuned version of Kuongan/xlm-roberta-base-ptbr-finetuned on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1806
• F1: 0.7467
• Roc Auc: 0.8408
• Accuracy: 0.7793

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: 2e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
0.2183	1.0	220	0.1558	0.4223	0.6979	0.7487
0.1705	2.0	440	0.1878	0.4040	0.6743	0.6681
0.144	3.0	660	0.1541	0.5332	0.7379	0.7394
0.1192	4.0	880	0.1628	0.5517	0.7503	0.7241
0.1023	5.0	1100	0.1771	0.5721	0.7531	0.7105
0.0848	6.0	1320	0.1715	0.6414	0.8023	0.7267
0.0638	7.0	1540	0.1637	0.6117	0.7727	0.7436
0.0413	8.0	1760	0.1885	0.6471	0.8067	0.7156
0.0323	9.0	1980	0.1899	0.6777	0.8021	0.7275
0.0262	10.0	2200	0.1672	0.7041	0.8228	0.7530
0.0254	11.0	2420	0.1765	0.7270	0.8300	0.7623
0.0194	12.0	2640	0.1752	0.7106	0.8132	0.7632
0.0117	13.0	2860	0.1821	0.7159	0.8297	0.7581
0.0086	14.0	3080	0.1761	0.7285	0.8374	0.7725
0.0106	15.0	3300	0.1787	0.7329	0.8356	0.7818
0.0087	16.0	3520	0.1802	0.7346	0.8369	0.7750
0.0101	17.0	3740	0.1826	0.7394	0.8426	0.7708
0.0107	18.0	3960	0.1806	0.7467	0.8408	0.7793
0.0078	19.0	4180	0.1808	0.7428	0.8414	0.7784
0.0071	20.0	4400	0.1806	0.7434	0.8416	0.7784

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0