bert-base-german-cased-own-data-ner
This model is a fine-tuned version of bert-base-german-cased on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.0535
- Precision: 0.7134
- Recall: 0.8536
- F1: 0.7772
- Accuracy: 0.9895
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 | 0.8 | 32 | 0.0308 | 0.7593 | 0.8 | 0.7791 | 0.9917 |
| No log | 1.6 | 64 | 0.0342 | 0.7756 | 0.8393 | 0.8062 | 0.9911 |
| No log | 2.4 | 96 | 0.0457 | 0.7764 | 0.8679 | 0.8196 | 0.9906 |
| No log | 3.2 | 128 | 0.0383 | 0.7524 | 0.8464 | 0.7966 | 0.9911 |
| No log | 4.0 | 160 | 0.0420 | 0.7539 | 0.8536 | 0.8007 | 0.9907 |
| No log | 4.8 | 192 | 0.0535 | 0.7134 | 0.8536 | 0.7772 | 0.9895 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.9.0+cu111
- Datasets 2.1.0
- Tokenizers 0.12.1