apriadiazriel/bert_base_ncbi

This model is a fine-tuned version of bert-base-uncased on the NCBI disease dataset. It achieves the following results on the evaluation set:

• Train Loss: 0.0168
• Validation Loss: 0.0518
• Precision: 0.8
• Recall: 0.8640
• F1: 0.8308
• Accuracy: 0.9860
• Epoch: 9

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:

• optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'module': 'keras.optimizers.schedules', 'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': 1017, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered_name': None}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.01}
• training_precision: float32

Training results

Train Loss	Validation Loss	Precision	Recall	F1	Accuracy	Epoch
0.1130	0.0547	0.7364	0.7916	0.7630	0.9832	0
0.0335	0.0497	0.7836	0.8513	0.8161	0.9850	1
0.0213	0.0518	0.8	0.8640	0.8308	0.9860	2
0.0166	0.0518	0.8	0.8640	0.8308	0.9860	3
0.0173	0.0518	0.8	0.8640	0.8308	0.9860	4
0.0174	0.0518	0.8	0.8640	0.8308	0.9860	5
0.0168	0.0518	0.8	0.8640	0.8308	0.9860	6
0.0172	0.0518	0.8	0.8640	0.8308	0.9860	7
0.0167	0.0518	0.8	0.8640	0.8308	0.9860	8
0.0168	0.0518	0.8	0.8640	0.8308	0.9860	9

Framework versions

• Transformers 4.48.3
• TensorFlow 2.18.0
• Datasets 3.3.1
• Tokenizers 0.21.0