metadata
license: mit
base_model: xlm-roberta-large
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
model-index:
- name: fine-tuned-KoreanNLI-KorNLI-with-xlm-roberta-large
results: []
fine-tuned-KoreanNLI-KorNLI-with-xlm-roberta-large
This model is a fine-tuned version of xlm-roberta-large on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.4428
- Accuracy: 0.8439
- F1: 0.8445
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: 1e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 128
- 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 | Accuracy | F1 |
|---|---|---|---|---|---|
| 0.4595 | 0.5 | 3654 | 0.4630 | 0.8064 | 0.8089 |
| 0.4138 | 1.0 | 7308 | 0.4497 | 0.8146 | 0.8165 |
| 0.3748 | 1.5 | 10962 | 0.4280 | 0.8420 | 0.8422 |
| 0.3687 | 2.0 | 14616 | 0.4161 | 0.8363 | 0.8376 |
| 0.3265 | 2.5 | 18270 | 0.4209 | 0.8459 | 0.8465 |
| 0.3392 | 3.0 | 21924 | 0.4107 | 0.8459 | 0.8453 |
| 0.2928 | 3.5 | 25578 | 0.4479 | 0.8395 | 0.8401 |
| 0.2975 | 4.0 | 29232 | 0.4428 | 0.8439 | 0.8445 |
Framework versions
- Transformers 4.31.0
- Pytorch 1.13.1
- Datasets 2.14.4
- Tokenizers 0.13.3