metadata
license: cc-by-nc-sa-4.0
tags:
- generated_from_trainer
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: test-finetuned-ner
results: []
test-finetuned-ner
This model is a fine-tuned version of hfl/chinese-pert-large on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.1687
- Precision: 0.7449
- Recall: 0.7717
- F1: 0.7581
- Accuracy: 0.9546
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: 3
- eval_batch_size: 3
- 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 |
|---|---|---|---|---|---|---|---|
| 0.1234 | 1.0 | 9387 | 0.1513 | 0.6954 | 0.7365 | 0.7153 | 0.9505 |
| 0.0841 | 2.0 | 18774 | 0.1462 | 0.7248 | 0.7630 | 0.7434 | 0.9533 |
| 0.0542 | 3.0 | 28161 | 0.1687 | 0.7449 | 0.7717 | 0.7581 | 0.9546 |
Framework versions
- Transformers 4.13.0
- Pytorch 1.8.0+cu111
- Datasets 2.4.0
- Tokenizers 0.10.3