ModernBERT-base-2-contract-sections-classification-v4-50-max

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.4092
Accuracy Evaluate: 0.9244
Precision Evaluate: 0.9280
Recall Evaluate: 0.9266
F1 Evaluate: 0.9265
Accuracy Sklearn: 0.9244
Precision Sklearn: 0.9252
Recall Sklearn: 0.9244
F1 Sklearn: 0.9239
Acuracia Rotulo Objeto: 0.9563
Acuracia Rotulo Obrigacoes: 0.9496
Acuracia Rotulo Valor: 0.8311
Acuracia Rotulo Vigencia: 0.9792
Acuracia Rotulo Rescisao: 0.9441
Acuracia Rotulo Foro: 0.9048
Acuracia Rotulo Reajuste: 0.8922
Acuracia Rotulo Fiscalizacao: 0.8485
Acuracia Rotulo Publicacao: 0.9885
Acuracia Rotulo Pagamento: 0.8829
Acuracia Rotulo Casos Omissos: 0.9103
Acuracia Rotulo Sancoes: 0.9722
Acuracia Rotulo Dotacao Orcamentaria: 0.9863

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: 2
eval_batch_size: 1
seed: 42
gradient_accumulation_steps: 16
total_train_batch_size: 32
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy Evaluate	Precision Evaluate	Recall Evaluate	F1 Evaluate	Accuracy Sklearn	Precision Sklearn	Recall Sklearn	F1 Sklearn	Acuracia Rotulo Objeto	Acuracia Rotulo Obrigacoes	Acuracia Rotulo Valor	Acuracia Rotulo Vigencia	Acuracia Rotulo Rescisao	Acuracia Rotulo Foro	Acuracia Rotulo Reajuste	Acuracia Rotulo Fiscalizacao	Acuracia Rotulo Publicacao	Acuracia Rotulo Pagamento	Acuracia Rotulo Casos Omissos	Acuracia Rotulo Sancoes	Acuracia Rotulo Dotacao Orcamentaria
No log	1.0	250	0.4141	0.9231	0.9252	0.9261	0.9254	0.9231	0.9231	0.9231	0.9229	0.9454	0.9419	0.8514	0.9792	0.9371	0.9048	0.8824	0.8409	1.0	0.9009	0.8974	0.9722	0.9863
0.357	2.0	500	0.4528	0.9194	0.9218	0.9253	0.9228	0.9194	0.9207	0.9194	0.9192	0.9508	0.9147	0.8311	0.9792	0.9371	0.9048	0.9020	0.8485	1.0	0.8919	0.9103	0.9722	0.9863
0.357	3.0	750	0.4269	0.925	0.9312	0.9270	0.9282	0.925	0.9262	0.925	0.9246	0.9617	0.9574	0.8243	0.9653	0.9441	0.9238	0.8824	0.8409	1.0	0.8829	0.9103	0.9722	0.9863
0.2319	4.0	1000	0.4197	0.9244	0.9283	0.9269	0.9267	0.9244	0.9252	0.9244	0.9238	0.9672	0.9457	0.8311	0.9792	0.9371	0.9048	0.8922	0.8409	1.0	0.8829	0.9103	0.9722	0.9863
0.2319	5.0	1250	0.4375	0.92	0.9191	0.9246	0.9209	0.92	0.9213	0.92	0.9198	0.9617	0.9225	0.8311	0.9722	0.9301	0.9048	0.8922	0.8561	0.9885	0.8919	0.9103	0.9722	0.9863
0.1568	6.0	1500	0.4203	0.9225	0.9232	0.9259	0.9239	0.9225	0.9232	0.9225	0.9222	0.9563	0.9380	0.8311	0.9792	0.9371	0.9048	0.8922	0.8485	0.9885	0.8919	0.9103	0.9722	0.9863
0.1568	7.0	1750	0.4056	0.9275	0.9294	0.9300	0.9289	0.9275	0.9280	0.9275	0.9270	0.9563	0.9419	0.8446	0.9861	0.9650	0.9048	0.8922	0.8409	0.9885	0.9009	0.9103	0.9722	0.9863
0.1107	8.0	2000	0.4097	0.9263	0.9289	0.9282	0.9278	0.9263	0.9267	0.9263	0.9257	0.9563	0.9535	0.8378	0.9861	0.9371	0.9143	0.8824	0.8409	0.9885	0.9009	0.9103	0.9722	0.9863
0.1107	9.0	2250	0.4176	0.9237	0.9266	0.9265	0.9259	0.9237	0.9244	0.9237	0.9233	0.9563	0.9457	0.8311	0.9792	0.9371	0.9048	0.9020	0.8485	0.9885	0.8829	0.9103	0.9722	0.9863
0.0923	10.0	2500	0.4092	0.9244	0.9280	0.9266	0.9265	0.9244	0.9252	0.9244	0.9239	0.9563	0.9496	0.8311	0.9792	0.9441	0.9048	0.8922	0.8485	0.9885	0.8829	0.9103	0.9722	0.9863

Framework versions

Transformers 4.49.0
Pytorch 2.6.0+cu124
Datasets 3.3.0
Tokenizers 0.21.0

marcelovidigal/ModernBERT-base-2-contract-sections-classification-v4-50-max

ModernBERT-base-2-contract-sections-classification-v4-50-max

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions