mongolian-ner-test-xlm-roberta-large-ner-hrl
This model is a fine-tuned version of bayartsogt/albert-mongolian on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.5337
- Precision: 0.3060
- Recall: 0.1406
- F1: 0.1927
- Accuracy: 0.8591
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: 16
- eval_batch_size: 32
- 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 |
|---|---|---|---|---|---|---|---|
| 0.6123 | 1.0 | 477 | 0.5570 | 0.2422 | 0.0999 | 0.1414 | 0.8536 |
| 0.5411 | 2.0 | 954 | 0.5407 | 0.2914 | 0.1294 | 0.1792 | 0.8572 |
| 0.5288 | 3.0 | 1431 | 0.5394 | 0.2944 | 0.1309 | 0.1812 | 0.8576 |
| 0.5212 | 4.0 | 1908 | 0.5346 | 0.3015 | 0.1324 | 0.1840 | 0.8581 |
| 0.5156 | 5.0 | 2385 | 0.5298 | 0.3131 | 0.1394 | 0.1929 | 0.8595 |
| 0.5103 | 6.0 | 2862 | 0.5301 | 0.3086 | 0.1419 | 0.1944 | 0.8595 |
| 0.5041 | 7.0 | 3339 | 0.5318 | 0.3083 | 0.1411 | 0.1936 | 0.8592 |
| 0.4981 | 8.0 | 3816 | 0.5308 | 0.3117 | 0.1421 | 0.1952 | 0.8595 |
| 0.4931 | 9.0 | 4293 | 0.5329 | 0.3062 | 0.1400 | 0.1922 | 0.8592 |
| 0.4885 | 10.0 | 4770 | 0.5337 | 0.3060 | 0.1406 | 0.1927 | 0.8591 |
Framework versions
- Transformers 4.28.1
- Pytorch 2.0.0+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3