metadata
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- article100v5_wikigold_split
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: Article_100v5_NER_Model_3Epochs_UNAUGMENTED
results:
- task:
name: Token Classification
type: token-classification
dataset:
name: article100v5_wikigold_split
type: article100v5_wikigold_split
args: default
metrics:
- name: Precision
type: precision
value: 0.024096385542168676
- name: Recall
type: recall
value: 0.0005104645227156713
- name: F1
type: f1
value: 0.000999750062484379
- name: Accuracy
type: accuracy
value: 0.7821558918567079
Article_100v5_NER_Model_3Epochs_UNAUGMENTED
This model is a fine-tuned version of bert-base-cased on the article100v5_wikigold_split dataset. It achieves the following results on the evaluation set:
- Loss: 0.5958
- Precision: 0.0241
- Recall: 0.0005
- F1: 0.0010
- Accuracy: 0.7822
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: 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: 3
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|---|---|---|---|---|---|---|---|
| No log | 1.0 | 13 | 0.7298 | 0.0 | 0.0 | 0.0 | 0.7816 |
| No log | 2.0 | 26 | 0.6272 | 0.0 | 0.0 | 0.0 | 0.7816 |
| No log | 3.0 | 39 | 0.5958 | 0.0241 | 0.0005 | 0.0010 | 0.7822 |
Framework versions
- Transformers 4.17.0
- Pytorch 1.11.0+cu113
- Datasets 2.4.0
- Tokenizers 0.11.6