Tagged_Uni_100v3_NER_Model_3Epochs_AUGMENTED

This model is a fine-tuned version of bert-base-cased on the tagged_uni100v3_wikigold_split dataset. It achieves the following results on the evaluation set:

• Loss: 0.4884
• Precision: 0.2764
• Recall: 0.1080
• F1: 0.1553
• Accuracy: 0.8106

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	26	0.6238	0.2	0.0089	0.0170	0.7822
No log	2.0	52	0.5210	0.2497	0.0587	0.0950	0.7971
No log	3.0	78	0.4884	0.2764	0.1080	0.1553	0.8106

Framework versions

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