bert-base-german-cased-20000-ner-uncased

This model is a fine-tuned version of dbmdz/bert-base-german-uncased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0617
• Precision: 0.8871
• Recall: 0.9013
• F1: 0.8941
• Accuracy: 0.9848

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: 5e-05
• train_batch_size: 96
• eval_batch_size: 96
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	0.34	64	0.0573	0.8859	0.8526	0.8689	0.9837
No log	0.68	128	0.0654	0.8107	0.8957	0.8511	0.9808
No log	1.02	192	0.0531	0.8654	0.8846	0.8749	0.9842
No log	1.35	256	0.0467	0.8847	0.8853	0.8850	0.9857
No log	1.69	320	0.0466	0.9102	0.8883	0.8992	0.9864
No log	2.03	384	0.0467	0.8794	0.8951	0.8872	0.9854
No log	2.37	448	0.0520	0.8864	0.9001	0.8932	0.9851
0.0531	2.71	512	0.0549	0.8894	0.8877	0.8885	0.9854
0.0531	3.05	576	0.0534	0.8942	0.8920	0.8931	0.9857
0.0531	3.39	640	0.0526	0.8917	0.8994	0.8956	0.9856
0.0531	3.72	704	0.0576	0.9049	0.8976	0.9012	0.9857
0.0531	4.06	768	0.0700	0.8529	0.9229	0.8865	0.9830
0.0531	4.4	832	0.0657	0.8716	0.9167	0.8936	0.9840
0.0531	4.74	896	0.0617	0.8871	0.9013	0.8941	0.9848

Framework versions

• Transformers 4.33.2
• Pytorch 2.0.1+cu118
• Datasets 2.14.5
• Tokenizers 0.13.3