bert-base-swedish-cased-new-finetuned-augmentation

This model is a fine-tuned version of KBLab/bert-base-swedish-cased-new on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1896
• F1: 0.5215
• Roc Auc: 0.7532
• Accuracy: 0.6931

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
0.4332	1.0	70	0.3604	0.1123	0.5414	0.4513
0.2766	2.0	140	0.2395	0.3940	0.6846	0.6606
0.2381	3.0	210	0.2095	0.3914	0.6785	0.6606
0.2096	4.0	280	0.2080	0.4761	0.7219	0.6751
0.186	5.0	350	0.2017	0.4803	0.7216	0.6570
0.1801	6.0	420	0.1937	0.4888	0.7343	0.6823
0.1333	7.0	490	0.1935	0.4903	0.7354	0.6606
0.1128	8.0	560	0.1962	0.4930	0.7356	0.6823
0.1107	9.0	630	0.2039	0.5069	0.7467	0.6643
0.0909	10.0	700	0.1896	0.5215	0.7532	0.6931
0.0811	11.0	770	0.2059	0.5147	0.7571	0.6679
0.0762	12.0	840	0.1988	0.5052	0.7423	0.6715
0.0673	13.0	910	0.1984	0.5160	0.7416	0.6968
0.0482	14.0	980	0.2044	0.5050	0.7430	0.6679

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.0
• Datasets 3.0.1
• Tokenizers 0.20.0