RoBERTa-ext-large-crf-lora-chinese-finetuned-ner
This model is a fine-tuned version of hfl/chinese-roberta-wwm-ext-large on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.4056
- Precision: 0.4202
- Recall: 0.5916
- F1: 0.4914
- Accuracy: 0.9456
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: 4
- eval_batch_size: 8
- 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 |
|---|---|---|---|---|---|---|---|
| 1.3615 | 1.0 | 503 | 0.8081 | 0.1274 | 0.1568 | 0.1406 | 0.9028 |
| 0.702 | 2.0 | 1006 | 0.5824 | 0.2954 | 0.4194 | 0.3467 | 0.9261 |
| 0.5585 | 3.0 | 1509 | 0.5107 | 0.3305 | 0.4922 | 0.3955 | 0.9323 |
| 0.4959 | 4.0 | 2012 | 0.4654 | 0.3716 | 0.5274 | 0.4360 | 0.9377 |
| 0.4614 | 5.0 | 2515 | 0.4427 | 0.3880 | 0.5493 | 0.4548 | 0.9399 |
| 0.4381 | 6.0 | 3018 | 0.4292 | 0.3996 | 0.5657 | 0.4684 | 0.9420 |
| 0.4233 | 7.0 | 3521 | 0.4166 | 0.4111 | 0.5813 | 0.4816 | 0.9441 |
| 0.4128 | 8.0 | 4024 | 0.4124 | 0.4144 | 0.5879 | 0.4862 | 0.9448 |
| 0.4008 | 9.0 | 4527 | 0.4067 | 0.4194 | 0.5904 | 0.4904 | 0.9455 |
| 0.3983 | 10.0 | 5030 | 0.4056 | 0.4202 | 0.5916 | 0.4914 | 0.9456 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.0