xlm-roberta-base_latin_original_amh-esp-eng_train_loss
This model is a fine-tuned version of FacebookAI/xlm-roberta-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.0261
- Spearman Corr: 0.7589
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: 32
- eval_batch_size: 128
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Spearman Corr |
---|---|---|---|---|
No log | 1.59 | 200 | 0.0183 | 0.7592 |
0.0498 | 3.17 | 400 | 0.0179 | 0.7531 |
0.0213 | 4.76 | 600 | 0.0205 | 0.7603 |
0.0161 | 6.35 | 800 | 0.0199 | 0.7582 |
0.0161 | 7.94 | 1000 | 0.0209 | 0.7641 |
0.0127 | 9.52 | 1200 | 0.0211 | 0.7647 |
0.0103 | 11.11 | 1400 | 0.0258 | 0.7592 |
0.0082 | 12.7 | 1600 | 0.0231 | 0.7571 |
0.007 | 14.29 | 1800 | 0.0251 | 0.7569 |
0.007 | 15.87 | 2000 | 0.0261 | 0.7589 |
Framework versions
- Transformers 4.37.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.0
- Tokenizers 0.15.2
- Downloads last month
- 8