NER_harem_bert-base-portuguese-cased

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

• Loss: 0.2351
• Precision: 0.6853
• Recall: 0.7378
• F1: 0.7106
• Accuracy: 0.9522

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: 3e-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: 300

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	16	0.7692	0.0	0.0	0.0	0.8358
No log	2.0	32	0.4831	0.3140	0.2731	0.2921	0.8790
No log	3.0	48	0.3405	0.4692	0.4897	0.4793	0.9119
No log	4.0	64	0.2747	0.5481	0.6156	0.5799	0.9340
No log	5.0	80	0.2282	0.6077	0.6758	0.6399	0.9443
No log	6.0	96	0.2145	0.6267	0.6892	0.6565	0.9479
No log	7.0	112	0.2223	0.6395	0.6926	0.6650	0.9493
No log	8.0	128	0.2100	0.6822	0.7378	0.7089	0.9530
No log	9.0	144	0.2077	0.6810	0.7497	0.7137	0.9537
No log	10.0	160	0.2173	0.6846	0.7460	0.7140	0.9523
No log	11.0	176	0.2226	0.7001	0.7594	0.7285	0.9542
No log	12.0	192	0.2204	0.7015	0.7568	0.7281	0.9538
No log	13.0	208	0.2278	0.6746	0.7411	0.7063	0.9533
No log	14.0	224	0.2351	0.6853	0.7378	0.7106	0.9522

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2