File size: 3,755 Bytes
c001272 869fe62 51d1934 c001272 869fe62 c001272 869fe62 c001272 869fe62 c001272 869fe62 c001272 869fe62 c001272 869fe62 c001272 869fe62 c001272 869fe62 c001272 51d1934 c001272 51d1934 c001272 51d1934 c001272 869fe62 c001272 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 |
---
license: apache-2.0
tags:
- generated_from_trainer
language: en
widget:
- text: "My name is Scott and I live in Columbus."
- text: "Apple was founded in 1976 by Steve Jobs, Steve Wozniak and Ronald Wayne."
datasets:
- conll2003
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: bert-large-uncased-finetuned-ner
results:
- task:
name: Token Classification
type: token-classification
dataset:
name: conll2003
type: conll2003
args: conll2003
metrics:
- name: Precision
type: precision
value: 0.9504719600222099
- name: Recall
type: recall
value: 0.9574896520863632
- name: F1
type: f1
value: 0.9539679001337494
- name: Accuracy
type: accuracy
value: 0.9885618059637473
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bert-large-uncased-finetuned-ner
This model is a fine-tuned version of [bert-large-uncased](https://huggingface.co./bert-large-uncased) on the conll2003 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0778
- Precision: 0.9505
- Recall: 0.9575
- F1: 0.9540
- Accuracy: 0.9886
## Model description
More information needed
#### Limitations and bias
This model is limited by its training dataset of entity-annotated news articles from a specific span of time. This may not generalize well for all use cases in different domains. Furthermore, the model occassionally tags subword tokens as entities and post-processing of results may be necessary to handle those cases.
#### How to use
You can use this model with Transformers *pipeline* for NER.
```python
from transformers import pipeline
from transformers import AutoTokenizer, AutoModelForTokenClassification
tokenizer = AutoTokenizer.from_pretrained("Jorgeutd/bert-large-uncased-finetuned-ner")
model = AutoModelForTokenClassification.from_pretrained("Jorgeutd/bert-large-uncased-finetuned-ner")
nlp = pipeline("ner", model=model, tokenizer=tokenizer)
example = "My name is Scott and I live in Ohio"
ner_results = nlp(example)
print(ner_results)
```
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 64
- 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 |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
| 0.1997 | 1.0 | 878 | 0.0576 | 0.9316 | 0.9257 | 0.9286 | 0.9837 |
| 0.04 | 2.0 | 1756 | 0.0490 | 0.9400 | 0.9513 | 0.9456 | 0.9870 |
| 0.0199 | 3.0 | 2634 | 0.0557 | 0.9436 | 0.9540 | 0.9488 | 0.9879 |
| 0.0112 | 4.0 | 3512 | 0.0602 | 0.9443 | 0.9569 | 0.9506 | 0.9881 |
| 0.0068 | 5.0 | 4390 | 0.0631 | 0.9451 | 0.9589 | 0.9520 | 0.9882 |
| 0.0044 | 6.0 | 5268 | 0.0638 | 0.9510 | 0.9567 | 0.9538 | 0.9885 |
| 0.003 | 7.0 | 6146 | 0.0722 | 0.9495 | 0.9560 | 0.9527 | 0.9885 |
| 0.0016 | 8.0 | 7024 | 0.0762 | 0.9491 | 0.9595 | 0.9543 | 0.9887 |
| 0.0018 | 9.0 | 7902 | 0.0769 | 0.9496 | 0.9542 | 0.9519 | 0.9883 |
| 0.0009 | 10.0 | 8780 | 0.0778 | 0.9505 | 0.9575 | 0.9540 | 0.9886 |
### Framework versions
- Transformers 4.16.2
- Pytorch 1.8.1+cu111
- Datasets 1.18.3
- Tokenizers 0.11.0
|