herbert-ner-finetuned

This model is a fine-tuned version of pczarnik/herbert-base-ner on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0014
• Precision: 0.9977
• Recall: 0.9982
• F1: 0.9980

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 327
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1
0.0072	0.7634	500	0.0061	0.9697	0.9905	0.98
0.0034	1.5267	1000	0.0031	0.9879	0.9960	0.9920
0.0016	2.2901	1500	0.0017	0.9958	0.9980	0.9969
0.0012	3.0534	2000	0.0017	0.9958	0.9975	0.9966
0.0007	3.8168	2500	0.0015	0.9976	0.9983	0.9980
0.0004	4.5802	3000	0.0015	0.9976	0.9977	0.9977
0.0006	5.3435	3500	0.0014	0.9977	0.9982	0.9980

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1
• Datasets 2.21.0
• Tokenizers 0.19.1