ABL_trad_j

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: 2.6432
• Accuracy: 0.6883
• F1: 0.6865

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-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: 32

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.9532	1.0	1500	0.9116	0.5825	0.5793
0.8601	2.0	3000	0.8433	0.6033	0.6016
0.7962	3.0	4500	0.8150	0.6275	0.6252
0.7633	4.0	6000	0.7969	0.635	0.6334
0.7153	5.0	7500	0.7825	0.6492	0.6483
0.678	6.0	9000	0.7910	0.6408	0.6392
0.6336	7.0	10500	0.7772	0.6608	0.6606
0.5981	8.0	12000	0.7863	0.6617	0.6605
0.5455	9.0	13500	0.7954	0.6658	0.6657
0.4972	10.0	15000	0.8206	0.6633	0.6623
0.4823	11.0	16500	0.8442	0.6683	0.6673
0.4258	12.0	18000	0.8966	0.6742	0.6734
0.4182	13.0	19500	0.9327	0.6767	0.6761
0.3588	14.0	21000	0.9780	0.6717	0.6689
0.3576	15.0	22500	1.0288	0.6833	0.6828
0.3252	16.0	24000	1.0873	0.6842	0.6836
0.3104	17.0	25500	1.1417	0.685	0.6847
0.2691	18.0	27000	1.2447	0.6842	0.6827
0.2559	19.0	28500	1.3480	0.6825	0.6816
0.2522	20.0	30000	1.4782	0.6867	0.6859
0.2234	21.0	31500	1.5748	0.6833	0.6815
0.1954	22.0	33000	1.7041	0.69	0.6897
0.1979	23.0	34500	1.8398	0.6808	0.6789
0.176	24.0	36000	1.9141	0.6867	0.6860
0.1862	25.0	37500	2.0105	0.6883	0.6881
0.1409	26.0	39000	2.1345	0.685	0.6840
0.1527	27.0	40500	2.2039	0.6858	0.6853
0.1474	28.0	42000	2.2990	0.6933	0.6920
0.1428	29.0	43500	2.3780	0.6883	0.6878
0.1348	30.0	45000	2.4859	0.6858	0.6839
0.1046	31.0	46500	2.5546	0.6825	0.6801
0.1147	32.0	48000	2.6432	0.6883	0.6865

Framework versions

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