distilbert-base-uncased-finetuned-ner
This model is a fine-tuned version of distilbert-base-uncased on the conll2003 dataset. It achieves the following results on the evaluation set:
- Loss: 0.0626
- Precision: 0.9248
- Recall: 0.9327
- F1: 0.9287
- Accuracy: 0.9832
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
- num_epochs: 3
Training results
Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
---|---|---|---|---|---|---|---|
0.252 | 1.0 | 878 | 0.0741 | 0.8929 | 0.9210 | 0.9067 | 0.9786 |
0.0516 | 2.0 | 1756 | 0.0606 | 0.9184 | 0.9295 | 0.9239 | 0.9823 |
0.0311 | 3.0 | 2634 | 0.0626 | 0.9248 | 0.9327 | 0.9287 | 0.9832 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.2+cu118
- Datasets 2.12.0
- Tokenizers 0.13.2
- Downloads last month
- 10
Finetuned from
Dataset used to train YijingZZZ/distilbert-base-uncased-finetuned-ner
Evaluation results
- Precision on conll2003validation set self-reported0.925
- Recall on conll2003validation set self-reported0.933
- F1 on conll2003validation set self-reported0.929
- Accuracy on conll2003validation set self-reported0.983