roberta-base_LeNER-Br

This model is a fine-tuned version of FacebookAI/roberta-base on the lener_br dataset. It achieves the following results on the evaluation set:

• Loss: nan
• Precision: 0.765
• Recall: 0.8416
• F1: 0.8015
• Accuracy: 0.9712

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: 2e-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: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.293	1.0	979	nan	0.5758	0.7525	0.6524	0.9542
0.0596	2.0	1958	nan	0.6546	0.7987	0.7195	0.9534
0.0376	3.0	2937	nan	0.7366	0.8339	0.7822	0.9672
0.0256	4.0	3916	nan	0.6975	0.8042	0.7471	0.9627
0.0192	5.0	4895	nan	0.7173	0.8317	0.7702	0.9646
0.013	6.0	5874	nan	0.7271	0.8498	0.7837	0.9605
0.013	7.0	6853	nan	0.7426	0.8537	0.7943	0.9680
0.0064	8.0	7832	nan	0.7493	0.8399	0.7920	0.9702
0.0052	9.0	8811	nan	0.7611	0.8273	0.7928	0.9725
0.0044	10.0	9790	nan	0.765	0.8416	0.8015	0.9712

Testing results

metrics={'test_loss': 0.08161260932683945, 'test_precision': 0.8342714196372732, 'test_recall': 0.8840291583830351, 'test_f1': 0.8584298584298585, 'test_accuracy': 0.9863512377202157, 'test_runtime': 20.4317, 'test_samples_per_second': 68.032, 'test_steps_per_second': 8.516})

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1