xlm-roberta-base_afr_corr_5e-06
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: 0.0220
- Spearman Corr: 0.7641
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: 5e-06
- 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 | 0.85 | 200 | 0.0322 | 0.6463 |
No log | 1.69 | 400 | 0.0245 | 0.7044 |
0.0562 | 2.54 | 600 | 0.0259 | 0.7112 |
0.0562 | 3.38 | 800 | 0.0239 | 0.7265 |
0.0271 | 4.23 | 1000 | 0.0240 | 0.7442 |
0.0271 | 5.07 | 1200 | 0.0228 | 0.7573 |
0.0271 | 5.92 | 1400 | 0.0231 | 0.7572 |
0.0229 | 6.77 | 1600 | 0.0227 | 0.7518 |
0.0229 | 7.61 | 1800 | 0.0218 | 0.7657 |
0.0205 | 8.46 | 2000 | 0.0220 | 0.7629 |
0.0205 | 9.3 | 2200 | 0.0213 | 0.7704 |
0.0186 | 10.15 | 2400 | 0.0222 | 0.7669 |
0.0186 | 10.99 | 2600 | 0.0226 | 0.7696 |
0.0186 | 11.84 | 2800 | 0.0218 | 0.7697 |
0.0168 | 12.68 | 3000 | 0.0217 | 0.7679 |
0.0168 | 13.53 | 3200 | 0.0235 | 0.7619 |
0.0153 | 14.38 | 3400 | 0.0221 | 0.7696 |
0.0153 | 15.22 | 3600 | 0.0225 | 0.7650 |
0.0144 | 16.07 | 3800 | 0.0220 | 0.7641 |
Framework versions
- Transformers 4.37.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.0
- Tokenizers 0.15.2
- Downloads last month
- 3