foxy-nlp-br

This model is a fine-tuned version of neuralmind/bert-large-portuguese-cased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1638
• Accuracy: 0.9714
• F1 Weighted: 0.9714
• F1 Macro: 0.9740
• F1 Saudacao: 1.0
• F1 Cancelamento: 1.0
• F1 Reclamacao: 0.8824
• F1 Financeiro: 1.0
• F1 Suporte Tecnico: 0.9545
• F1 Elogio: 0.9412
• F1 Informacao: 1.0
• F1 Pedido Entrega: 1.0
• F1 Conta Perfil: 0.9615
• F1 Negociacao Retencao: 1.0
• F1 Min Class: 0.8824

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: 4.197342969974621e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 0.0022118889184775006
• num_epochs: 22

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Weighted	F1 Macro	F1 Saudacao	F1 Cancelamento	F1 Reclamacao	F1 Financeiro	F1 Suporte Tecnico	F1 Elogio	F1 Informacao	F1 Pedido Entrega	F1 Conta Perfil	F1 Negociacao Retencao	F1 Min Class
0.8071	1.0	125	0.8255	0.7943	0.7724	0.7549	0.8966	0.7059	0.7692	0.7619	0.8261	0.9143	0.75	0.9375	0.8333	0.1538	0.1538
0.2542	2.0	250	0.3204	0.9257	0.9254	0.9301	0.9630	0.9714	0.8824	1.0	0.8837	0.9412	0.8421	0.9375	0.9231	0.9565	0.8421
0.0880	3.0	375	0.2739	0.9314	0.9306	0.9334	0.9286	1.0	0.8824	1.0	0.9048	0.9412	0.8571	0.9375	0.9259	0.9565	0.8571
0.0176	4.0	500	0.2407	0.9486	0.9482	0.9519	1.0	1.0	0.8824	1.0	0.9048	0.9412	0.9231	0.9677	0.9434	0.9565	0.8824
0.0129	5.0	625	0.2209	0.9543	0.9541	0.9563	1.0	1.0	0.8824	1.0	0.9302	0.9412	0.9231	0.9677	0.9615	0.9565	0.8824
0.0072	6.0	750	0.2345	0.9543	0.9541	0.9569	1.0	1.0	0.8824	1.0	0.9302	0.9412	0.9474	0.9677	0.9434	0.9565	0.8824
0.0085	7.0	875	0.2161	0.96	0.9599	0.9638	1.0	1.0	0.8824	1.0	0.9302	0.9412	0.9730	0.9677	0.9434	1.0	0.8824
0.0061	8.0	1000	0.2211	0.96	0.9599	0.9638	1.0	1.0	0.8824	1.0	0.9302	0.9412	0.9730	0.9677	0.9434	1.0	0.8824
0.0076	9.0	1125	0.1966	0.96	0.9599	0.9638	1.0	1.0	0.8824	1.0	0.9302	0.9412	0.9730	0.9677	0.9434	1.0	0.8824
0.0053	10.0	1250	0.1642	0.9714	0.9714	0.9740	1.0	1.0	0.8824	1.0	0.9545	0.9412	1.0	1.0	0.9615	1.0	0.8824
0.0069	11.0	1375	0.1517	0.9657	0.9657	0.9680	1.0	1.0	0.8824	1.0	0.9545	0.9412	0.9730	0.9677	0.9615	1.0	0.8824
0.0049	12.0	1500	0.1775	0.96	0.9599	0.9638	1.0	1.0	0.8824	1.0	0.9302	0.9412	0.9730	0.9677	0.9434	1.0	0.8824
0.0051	13.0	1625	0.1730	0.9657	0.9657	0.9680	1.0	1.0	0.8824	1.0	0.9545	0.9412	0.9730	0.9677	0.9615	1.0	0.8824

Framework versions

• Transformers 5.3.0
• Pytorch 2.10.0+cu128
• Datasets 3.0.0
• Tokenizers 0.22.2