metadata
license: mit
base_model: FacebookAI/xlm-roberta-base
tags:
- generated_from_trainer
model-index:
- name: xlm-roberta-base_latin_kin-amh-eng_train_loss
results: []
xlm-roberta-base_latin_kin-amh-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: 0.0326
- Spearman Corr: 0.7395
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 | 0.59 | 200 | 0.0303 | 0.6125 |
| No log | 1.17 | 400 | 0.0269 | 0.6780 |
| No log | 1.76 | 600 | 0.0393 | 0.6855 |
| 0.036 | 2.35 | 800 | 0.0338 | 0.7111 |
| 0.036 | 2.93 | 1000 | 0.0303 | 0.6886 |
| 0.036 | 3.52 | 1200 | 0.0327 | 0.7025 |
| 0.0243 | 4.11 | 1400 | 0.0269 | 0.7220 |
| 0.0243 | 4.69 | 1600 | 0.0287 | 0.7246 |
| 0.0243 | 5.28 | 1800 | 0.0260 | 0.7336 |
| 0.0243 | 5.87 | 2000 | 0.0266 | 0.7234 |
| 0.0185 | 6.45 | 2200 | 0.0252 | 0.7347 |
| 0.0185 | 7.04 | 2400 | 0.0281 | 0.7276 |
| 0.0185 | 7.62 | 2600 | 0.0294 | 0.7298 |
| 0.0141 | 8.21 | 2800 | 0.0274 | 0.7219 |
| 0.0141 | 8.8 | 3000 | 0.0285 | 0.7260 |
| 0.0141 | 9.38 | 3200 | 0.0276 | 0.7315 |
| 0.0141 | 9.97 | 3400 | 0.0291 | 0.7329 |
| 0.0109 | 10.56 | 3600 | 0.0310 | 0.7339 |
| 0.0109 | 11.14 | 3800 | 0.0326 | 0.7395 |
Framework versions
- Transformers 4.37.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.0
- Tokenizers 0.15.2