metadata

license: mit
tags:
  - generated_from_trainer
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: bert-base-german-cased-20000-ner
    results: []

bert-base-german-cased-20000-ner

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

Loss: 0.0368
Precision: 0.8221
Recall: 0.875
F1: 0.8478
Accuracy: 0.9920

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: 32
eval_batch_size: 32
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	0.1	64	0.0427	0.7796	0.8714	0.8229	0.9893
No log	0.19	128	0.0472	0.5471	0.85	0.6657	0.9831
No log	0.29	192	0.0384	0.7897	0.8179	0.8035	0.9899
No log	0.38	256	0.0488	0.4970	0.8786	0.6348	0.9793
No log	0.48	320	0.0412	0.7548	0.8464	0.7980	0.9895
No log	0.58	384	0.0437	0.8373	0.8821	0.8591	0.9914
No log	0.67	448	0.0399	0.7727	0.85	0.8095	0.9899
0.0914	0.77	512	0.0394	0.7859	0.8786	0.8297	0.9899
0.0914	0.86	576	0.0368	0.8221	0.875	0.8478	0.9920

Framework versions

Transformers 4.18.0
Pytorch 1.9.0+cu111
Datasets 2.1.0
Tokenizers 0.12.1