ABL_trad_k

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

• Loss: 2.8641
• Accuracy: 0.6842
• F1: 0.6826

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: 1e-06
• 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: 36

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.9411	1.0	1500	0.9017	0.5792	0.5764
0.8328	2.0	3000	0.8458	0.62	0.6188
0.8006	3.0	4500	0.8183	0.64	0.6391
0.7283	4.0	6000	0.8154	0.6442	0.6430
0.7006	5.0	7500	0.7978	0.6492	0.6487
0.6555	6.0	9000	0.8009	0.6542	0.6536
0.6263	7.0	10500	0.8033	0.6617	0.6612
0.5805	8.0	12000	0.8155	0.6658	0.6657
0.5385	9.0	13500	0.8608	0.675	0.6729
0.5108	10.0	15000	0.8545	0.6733	0.6732
0.4791	11.0	16500	0.8950	0.6758	0.6750
0.4423	12.0	18000	0.9145	0.6792	0.6790
0.4295	13.0	19500	0.9497	0.6708	0.6707
0.3782	14.0	21000	1.0309	0.6742	0.6734
0.3656	15.0	22500	1.0706	0.6783	0.6774
0.3312	16.0	24000	1.1327	0.6733	0.6730
0.3008	17.0	25500	1.1870	0.6825	0.6822
0.2851	18.0	27000	1.3284	0.685	0.6847
0.2636	19.0	28500	1.4260	0.6858	0.6851
0.2752	20.0	30000	1.4733	0.6833	0.6833
0.2231	21.0	31500	1.6163	0.68	0.6800
0.2052	22.0	33000	1.7674	0.6792	0.6786
0.198	23.0	34500	1.8474	0.6833	0.6827
0.1854	24.0	36000	1.9509	0.6775	0.6766
0.1944	25.0	37500	2.0660	0.68	0.6790
0.1649	26.0	39000	2.1718	0.6825	0.6812
0.1443	27.0	40500	2.3664	0.68	0.6791
0.1251	28.0	42000	2.4144	0.6833	0.6827
0.1357	29.0	43500	2.4407	0.6875	0.6875
0.1279	30.0	45000	2.4419	0.6933	0.6932
0.1112	31.0	46500	2.5989	0.6833	0.6828
0.098	32.0	48000	2.6390	0.68	0.6792
0.0864	33.0	49500	2.7293	0.6792	0.6780
0.0897	34.0	51000	2.7814	0.6833	0.6820
0.0869	35.0	52500	2.8468	0.68	0.6787
0.0834	36.0	54000	2.8641	0.6842	0.6826

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.1