CoNLL2003_NER_BERT_Base_Cased

This model is a fine-tuned version of bert-base-cased on the conll2003 dataset. It achieves the following results on the evaluation set:

• Loss: 0.0571
• Precision: 0.9250
• Recall: 0.9458
• F1: 0.9353
• Accuracy: 0.9855

Model description

BERT Base Cased Model available at: bert-base-cased

Intended uses & limitations

Named Entity Recognition Task (English)

Training and evaluation data

CONLL2023 Dataset available at: conll2003 dataset

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.298	1.0	220	0.0792	0.8463	0.8995	0.8721	0.9769
0.0647	2.0	440	0.0617	0.9088	0.9362	0.9223	0.9830
0.0394	3.0	660	0.0574	0.9207	0.9443	0.9324	0.9846
0.0286	4.0	880	0.0559	0.9195	0.9438	0.9315	0.9855
0.0222	5.0	1100	0.0571	0.9250	0.9458	0.9353	0.9855

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1