|
--- |
|
license: mit |
|
tags: |
|
- generated_from_trainer |
|
datasets: |
|
- tner/ontonotes5 |
|
metrics: |
|
- precision |
|
- recall |
|
- f1 |
|
- accuracy |
|
widget: |
|
- text: 'I am Jack. I live in California and I work at Apple ' |
|
example_title: Example 1 |
|
- text: 'Wow this book is amazing and costs only 4€ ' |
|
example_title: Example 2 |
|
base_model: distilbert-base-cased |
|
model-index: |
|
- name: distilbert-finetuned-ner-ontonotes |
|
results: |
|
- task: |
|
type: token-classification |
|
name: Token Classification |
|
dataset: |
|
name: ontonotes5 |
|
type: ontonotes5 |
|
config: ontonotes5 |
|
split: train |
|
args: ontonotes5 |
|
metrics: |
|
- type: precision |
|
value: 0.8535359959297889 |
|
name: Precision |
|
- type: recall |
|
value: 0.8788553467356427 |
|
name: Recall |
|
- type: f1 |
|
value: 0.8660106468785288 |
|
name: F1 |
|
- type: accuracy |
|
value: 0.9749625470373822 |
|
name: Accuracy |
|
--- |
|
|
|
<!-- 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. --> |
|
|
|
# distilbert-finetuned-ner-ontonotes |
|
|
|
This model is a fine-tuned version of [distilbert-base-cased](https://huggingface.co/distilbert-base-cased) on the ontonotes5 dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.1448 |
|
- Precision: 0.8535 |
|
- Recall: 0.8789 |
|
- F1: 0.8660 |
|
- Accuracy: 0.9750 |
|
|
|
## Model description |
|
|
|
Token classification experiment, NER, on business topics. |
|
|
|
## Intended uses & limitations |
|
|
|
The model can be used on token classification, in particular NER. It is fine tuned on business domain. |
|
|
|
## Training and evaluation data |
|
|
|
The dataset used is [ontonotes5](https://huggingface.co/datasets/tner/ontonotes5) |
|
|
|
## Training procedure |
|
|
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 2e-05 |
|
- train_batch_size: 8 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 6 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy | |
|
|:-------------:|:-----:|:-----:|:---------------:|:---------:|:------:|:------:|:--------:| |
|
| 0.0937 | 1.0 | 7491 | 0.0998 | 0.8367 | 0.8587 | 0.8475 | 0.9731 | |
|
| 0.0572 | 2.0 | 14982 | 0.1084 | 0.8338 | 0.8759 | 0.8543 | 0.9737 | |
|
| 0.0403 | 3.0 | 22473 | 0.1145 | 0.8521 | 0.8707 | 0.8613 | 0.9748 | |
|
| 0.0265 | 4.0 | 29964 | 0.1222 | 0.8535 | 0.8815 | 0.8672 | 0.9752 | |
|
| 0.0148 | 5.0 | 37455 | 0.1365 | 0.8536 | 0.8770 | 0.8651 | 0.9747 | |
|
| 0.0111 | 6.0 | 44946 | 0.1448 | 0.8535 | 0.8789 | 0.8660 | 0.9750 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.22.1 |
|
- Pytorch 1.12.1+cu113 |
|
- Datasets 2.5.1 |
|
- Tokenizers 0.12.1 |
|
|
