chinese-roberta-wwm-ext-large-lora-ner
This model is a fine-tuned version of hfl/chinese-roberta-wwm-ext-large on the gyr66/privacy_detection dataset. It achieves the following results on the evaluation set:
- Loss: 0.3302
- Precision: 0.6010
- Recall: 0.7258
- F1: 0.6575
- Accuracy: 0.9106
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: 0.001
- train_batch_size: 28
- eval_batch_size: 56
- 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 |
|---|---|---|---|---|---|---|---|
| 0.729 | 1.0 | 72 | 0.3562 | 0.4495 | 0.5818 | 0.5072 | 0.8865 |
| 0.3155 | 2.0 | 144 | 0.3243 | 0.5155 | 0.6636 | 0.5803 | 0.8972 |
| 0.2561 | 3.0 | 216 | 0.3021 | 0.5599 | 0.7004 | 0.6223 | 0.9067 |
| 0.2283 | 4.0 | 288 | 0.3049 | 0.5670 | 0.6984 | 0.6259 | 0.9044 |
| 0.1952 | 5.0 | 360 | 0.3144 | 0.5836 | 0.7145 | 0.6424 | 0.9076 |
| 0.174 | 6.0 | 432 | 0.3157 | 0.5787 | 0.7183 | 0.6410 | 0.9063 |
| 0.155 | 7.0 | 504 | 0.3223 | 0.5966 | 0.7246 | 0.6544 | 0.9083 |
| 0.1436 | 8.0 | 576 | 0.3267 | 0.5921 | 0.7210 | 0.6502 | 0.9088 |
| 0.1298 | 9.0 | 648 | 0.3345 | 0.5965 | 0.7276 | 0.6556 | 0.9089 |
| 0.1226 | 10.0 | 720 | 0.3302 | 0.6010 | 0.7258 | 0.6575 | 0.9106 |
Framework versions
- Transformers 4.27.3
- Pytorch 2.0.1+cu117
- Datasets 2.14.5
- Tokenizers 0.13.2