CS221-bert-base-uncased-finetuned-semeval-NTdeu

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

• Loss: 0.5016
• F1: 0.6321
• Roc Auc: 0.7571
• Accuracy: 0.5566

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: 8
• eval_batch_size: 8
• 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
0.3253	1.0	261	0.3549	0.3597	0.6101	0.4012
0.3116	2.0	522	0.3043	0.5547	0.7009	0.5144
0.2182	3.0	783	0.3106	0.5885	0.7273	0.5163
0.1216	4.0	1044	0.3249	0.5851	0.7324	0.5048
0.077	5.0	1305	0.3578	0.5931	0.7298	0.5202
0.0335	6.0	1566	0.3857	0.6011	0.7338	0.5298
0.0214	7.0	1827	0.4070	0.61	0.7469	0.5278
0.0092	8.0	2088	0.4319	0.6198	0.7540	0.5374
0.0081	9.0	2349	0.4407	0.6239	0.7534	0.5393
0.006	10.0	2610	0.4525	0.6236	0.7521	0.5470
0.004	11.0	2871	0.4609	0.6266	0.7524	0.5470
0.0036	12.0	3132	0.4741	0.6237	0.7498	0.5451
0.0026	13.0	3393	0.4845	0.6223	0.7500	0.5432
0.0033	14.0	3654	0.4889	0.6314	0.7575	0.5528
0.003	15.0	3915	0.4953	0.6264	0.7518	0.5528
0.0034	16.0	4176	0.4986	0.6288	0.7555	0.5528
0.0023	17.0	4437	0.5016	0.6321	0.7571	0.5566
0.0027	18.0	4698	0.5027	0.6232	0.7508	0.5470
0.0032	19.0	4959	0.5029	0.6302	0.7553	0.5528
0.0029	20.0	5220	0.5029	0.6294	0.7551	0.5509

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu121
• Tokenizers 0.21.0