README.md

metadata

license: mit
tags:
  - generated_from_trainer
model-index:
  - name: bert_base_tcm_teste
    results: []

bert_base_tcm_teste

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

• Loss: 0.0205
• Criterio Julgamento Precision: 0.7719
• Criterio Julgamento Recall: 0.8462
• Criterio Julgamento F1: 0.8073
• Criterio Julgamento Number: 104
• Data Sessao Precision: 0.7812
• Data Sessao Recall: 0.9091
• Data Sessao F1: 0.8403
• Data Sessao Number: 55
• Modalidade Licitacao Precision: 0.9507
• Modalidade Licitacao Recall: 0.9620
• Modalidade Licitacao F1: 0.9563
• Modalidade Licitacao Number: 421
• Numero Exercicio Precision: 0.9375
• Numero Exercicio Recall: 0.9730
• Numero Exercicio F1: 0.9549
• Numero Exercicio Number: 185
• Objeto Licitacao Precision: 0.5309
• Objeto Licitacao Recall: 0.7288
• Objeto Licitacao F1: 0.6143
• Objeto Licitacao Number: 59
• Valor Objeto Precision: 0.8409
• Valor Objeto Recall: 0.9024
• Valor Objeto F1: 0.8706
• Valor Objeto Number: 41
• Overall Precision: 0.8719
• Overall Recall: 0.9283
• Overall F1: 0.8992
• Overall Accuracy: 0.9967

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: 50.0

Training results

Training Loss	Epoch	Step	Validation Loss	Criterio Julgamento Precision	Criterio Julgamento Recall	Criterio Julgamento F1	Criterio Julgamento Number	Data Sessao Precision	Data Sessao Recall	Data Sessao F1	Data Sessao Number	Modalidade Licitacao Precision	Modalidade Licitacao Recall	Modalidade Licitacao F1	Modalidade Licitacao Number	Numero Exercicio Precision	Numero Exercicio Recall	Numero Exercicio F1	Numero Exercicio Number	Objeto Licitacao Precision	Objeto Licitacao Recall	Objeto Licitacao F1	Objeto Licitacao Number	Valor Objeto Precision	Valor Objeto Recall	Valor Objeto F1	Valor Objeto Number	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
0.0168	0.96	2750	0.0169	0.7016	0.8365	0.7632	104	0.6707	1.0	0.8029	55	0.9424	0.9715	0.9567	421	0.9110	0.9405	0.9255	185	0.3304	0.6271	0.4327	59	0.76	0.9268	0.8352	41	0.8056	0.9249	0.8611	0.9950
0.0164	1.92	5500	0.0125	0.7565	0.8365	0.7945	104	0.6923	0.9818	0.8120	55	0.9491	0.9739	0.9613	421	0.9375	0.9730	0.9549	185	0.4138	0.6102	0.4932	59	0.8085	0.9268	0.8636	41	0.8465	0.9306	0.8866	0.9965
0.0076	2.88	8250	0.0204	0.7184	0.7115	0.7150	104	0.8070	0.8364	0.8214	55	0.9468	0.9715	0.9590	421	0.9282	0.9784	0.9526	185	0.4783	0.5593	0.5156	59	0.7209	0.7561	0.7381	41	0.8610	0.8948	0.8776	0.9961
0.0067	3.84	11000	0.0168	0.7589	0.8173	0.7870	104	0.8	0.8	0.8000	55	0.9487	0.9667	0.9576	421	0.9319	0.9622	0.9468	185	0.5309	0.7288	0.6143	59	0.8636	0.9268	0.8941	41	0.8717	0.9191	0.8948	0.9965
0.0043	4.8	13750	0.0144	0.736	0.8846	0.8035	104	0.8033	0.8909	0.8448	55	0.9512	0.9715	0.9612	421	0.9316	0.9568	0.944	185	0.5135	0.6441	0.5714	59	0.8444	0.9268	0.8837	41	0.8681	0.9283	0.8972	0.9967
0.0072	5.76	16500	0.0161	0.8091	0.8558	0.8318	104	0.7237	1.0	0.8397	55	0.9487	0.9667	0.9576	421	0.9326	0.9730	0.9524	185	0.4318	0.6441	0.5170	59	0.8222	0.9024	0.8605	41	0.8565	0.9318	0.8926	0.9966
0.003	6.72	19250	0.0205	0.7719	0.8462	0.8073	104	0.7812	0.9091	0.8403	55	0.9507	0.9620	0.9563	421	0.9375	0.9730	0.9549	185	0.5309	0.7288	0.6143	59	0.8409	0.9024	0.8706	41	0.8719	0.9283	0.8992	0.9967
0.0033	7.68	22000	0.0197	0.7736	0.7885	0.7810	104	0.7463	0.9091	0.8197	55	0.9466	0.9691	0.9577	421	0.9227	0.9676	0.9446	185	0.5286	0.6271	0.5736	59	0.7442	0.7805	0.7619	41	0.8650	0.9110	0.8874	0.9964
0.0043	8.64	24750	0.0250	0.7607	0.8558	0.8054	104	0.7612	0.9273	0.8361	55	0.9400	0.9667	0.9532	421	0.9427	0.9784	0.9602	185	0.5479	0.6780	0.6061	59	0.8043	0.9024	0.8506	41	0.8675	0.9306	0.8979	0.9965
0.0014	9.61	27500	0.0257	0.8018	0.8558	0.8279	104	0.7391	0.9273	0.8226	55	0.9417	0.9596	0.9506	421	0.9372	0.9676	0.9521	185	0.5143	0.6102	0.5581	59	0.8	0.8780	0.8372	41	0.8689	0.9191	0.8933	0.9966
0.0025	10.57	30250	0.0258	0.7798	0.8173	0.7981	104	0.7424	0.8909	0.8099	55	0.9465	0.9667	0.9565	421	0.9424	0.9730	0.9574	185	0.5352	0.6441	0.5846	59	0.8222	0.9024	0.8605	41	0.8728	0.9202	0.8959	0.9963
0.0016	11.53	33000	0.0273	0.7925	0.8077	0.8000	104	0.7246	0.9091	0.8065	55	0.9485	0.9620	0.9552	421	0.9282	0.9784	0.9526	185	0.56	0.7119	0.6269	59	0.8409	0.9024	0.8706	41	0.8723	0.9237	0.8972	0.9964

Framework versions

• Transformers 4.21.0.dev0
• Pytorch 1.11.0+cu113
• Datasets 2.3.2
• Tokenizers 0.12.1