german-bert-finetuned-ner

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0907
• Precision: 0.8143
• Recall: 0.8180
• F1: 0.8161
• Accuracy: 0.9893

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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	120	0.0715	0.7092	0.8	0.7518	0.9881
No log	2.0	240	0.0935	0.6588	0.8157	0.7289	0.9845
No log	3.0	360	0.0664	0.7303	0.7910	0.7594	0.9885
No log	4.0	480	0.0647	0.6691	0.8225	0.7379	0.9872
0.0027	5.0	600	0.0752	0.8409	0.7955	0.8176	0.9900
0.0027	6.0	720	0.0658	0.7105	0.8382	0.7691	0.9887
0.0027	7.0	840	0.0818	0.8364	0.8045	0.8202	0.9896
0.0027	8.0	960	0.0791	0.7660	0.8315	0.7974	0.9892
0.0013	9.0	1080	0.0791	0.7730	0.8112	0.7917	0.9893
0.0013	10.0	1200	0.0809	0.8117	0.8135	0.8126	0.9889
0.0013	11.0	1320	0.0851	0.8085	0.8157	0.8121	0.9894
0.0013	12.0	1440	0.0875	0.8361	0.8022	0.8188	0.9894
0.0004	13.0	1560	0.0892	0.8395	0.8112	0.8251	0.9893
0.0004	14.0	1680	0.0857	0.7978	0.8337	0.8154	0.9894
0.0004	15.0	1800	0.0848	0.7931	0.8270	0.8097	0.9895
0.0004	16.0	1920	0.0867	0.8053	0.8180	0.8116	0.9896
0.0002	17.0	2040	0.0866	0.7842	0.8247	0.8039	0.9891
0.0002	18.0	2160	0.0885	0.7952	0.8202	0.8075	0.9893
0.0002	19.0	2280	0.0895	0.7948	0.8180	0.8062	0.9894
0.0002	20.0	2400	0.0907	0.8143	0.8180	0.8161	0.9893

Framework versions

• Transformers 4.21.2
• Pytorch 1.12.1+cu113
• Datasets 2.4.0
• Tokenizers 0.12.1