metadata

license: apache-2.0
base_model: google-bert/bert-base-multilingual-cased
tags:
  - generated_from_trainer
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: pii_mbert_az
    results: []

pii_mbert_az

This model is a fine-tuned version of google-bert/bert-base-multilingual-cased on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.1319
Precision: 0.8726
Recall: 0.9026
F1: 0.8874
Accuracy: 0.9619

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: 32
eval_batch_size: 32
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: reduce_lr_on_plateau
num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	313	0.1464	0.8797	0.8615	0.8705	0.9587
0.2128	2.0	626	0.1319	0.8726	0.9026	0.8874	0.9619
0.2128	3.0	939	0.1461	0.8689	0.8924	0.8805	0.9596
0.0783	4.0	1252	0.1529	0.8837	0.9049	0.8942	0.9620
0.0443	5.0	1565	0.1921	0.8657	0.9157	0.8900	0.9615
0.0443	6.0	1878	0.1647	0.8975	0.9224	0.9098	0.9685
0.0201	7.0	2191	0.1725	0.8904	0.9183	0.9041	0.9674
0.0098	8.0	2504	0.1766	0.8917	0.9199	0.9056	0.9682
0.0098	9.0	2817	0.1756	0.8926	0.9202	0.9062	0.9686
0.007	10.0	3130	0.1763	0.8916	0.9189	0.9051	0.9684
0.007	11.0	3443	0.1772	0.8907	0.9183	0.9043	0.9682
0.007	12.0	3756	0.1773	0.8895	0.9173	0.9032	0.9680
0.0067	13.0	4069	0.1775	0.8892	0.9170	0.9029	0.9680
0.0067	14.0	4382	0.1775	0.8897	0.9170	0.9032	0.9679
0.0062	15.0	4695	0.1775	0.8897	0.9170	0.9032	0.9679

Framework versions

Transformers 4.41.0
Pytorch 2.3.0+cu121
Datasets 2.19.1
Tokenizers 0.19.1