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: 0.0611
- Precision: 0.9230
- Recall: 0.9366
- F1: 0.9298
- 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 |
---|---|---|---|---|---|---|---|
0.2349 | 1.0 | 878 | 0.0736 | 0.9140 | 0.9211 | 0.9175 | 0.9803 |
0.0546 | 2.0 | 1756 | 0.0582 | 0.9244 | 0.9368 | 0.9305 | 0.9830 |
0.03 | 3.0 | 2634 | 0.0611 | 0.9230 | 0.9366 | 0.9298 | 0.9832 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.10.0+cu111
- Datasets 2.0.0
- Tokenizers 0.11.6
- Downloads last month
- 3
Inference Providers
NEW
This model is not currently available via any of the supported third-party Inference Providers, and
the model is not deployed on the HF Inference API.
Dataset used to train ACSHCSE/distilbert-base-uncased-finetuned-ner
Evaluation results
- Precision on conll2003self-reported0.923
- Recall on conll2003self-reported0.937
- F1 on conll2003self-reported0.930
- Accuracy on conll2003self-reported0.983