NER_harem_bert-large-portuguese-cased

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

• Loss: 0.2487
• Precision: 0.7077
• Recall: 0.7553
• F1: 0.7308
• Accuracy: 0.9551

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	0.6334	0.0163	0.0078	0.0106	0.8468
No log	2.0	32	0.4537	0.2614	0.3112	0.2841	0.8826
No log	3.0	48	0.3117	0.5262	0.5671	0.5458	0.9231
No log	4.0	64	0.2421	0.5852	0.6631	0.6217	0.9385
No log	5.0	80	0.2099	0.5950	0.6855	0.6370	0.9479
No log	6.0	96	0.2153	0.6810	0.7464	0.7122	0.9551
No log	7.0	112	0.2270	0.6894	0.7198	0.7043	0.9546
No log	8.0	128	0.2213	0.6918	0.7437	0.7168	0.9554
No log	9.0	144	0.2299	0.7021	0.7564	0.7283	0.9545
No log	10.0	160	0.2256	0.7002	0.7591	0.7284	0.9562
No log	11.0	176	0.2169	0.7100	0.7736	0.7404	0.9568
No log	12.0	192	0.2266	0.6981	0.7740	0.7341	0.9571
No log	13.0	208	0.2322	0.7093	0.7620	0.7347	0.9570
No log	14.0	224	0.2487	0.7077	0.7553	0.7308	0.9551

Framework versions

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