Tagged_Uni_500v6_NER_Model_3Epochs_AUGMENTED

This model is a fine-tuned version of bert-base-cased on the tagged_uni500v6_wikigold_split dataset. It achieves the following results on the evaluation set:

• Loss: 0.2386
• Precision: 0.6992
• Recall: 0.6987
• F1: 0.6989
• Accuracy: 0.9318

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: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	182	0.2452	0.5956	0.5432	0.5682	0.9189
No log	2.0	364	0.2571	0.6832	0.6354	0.6584	0.9204
0.1093	3.0	546	0.2386	0.6992	0.6987	0.6989	0.9318

Framework versions

• Transformers 4.17.0
• Pytorch 1.11.0+cu113
• Datasets 2.4.0
• Tokenizers 0.11.6