metadata
license: mit
base_model: FacebookAI/xlm-roberta-base
tags:
- generated_from_trainer
model-index:
- name: xlm-roberta-base_latin_original_kin-amh-eng_train_loss
results: []
xlm-roberta-base_latin_original_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: 0.0293
- Spearman Corr: 0.7416
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 | 1.75 | 200 | 0.0265 | 0.6827 |
| 0.0364 | 3.51 | 400 | 0.0242 | 0.7207 |
| 0.0205 | 5.26 | 600 | 0.0258 | 0.7479 |
| 0.0146 | 7.02 | 800 | 0.0410 | 0.7467 |
| 0.0117 | 8.77 | 1000 | 0.0300 | 0.7324 |
| 0.0092 | 10.53 | 1200 | 0.0282 | 0.7313 |
| 0.0078 | 12.28 | 1400 | 0.0299 | 0.7357 |
| 0.0067 | 14.04 | 1600 | 0.0308 | 0.7329 |
| 0.0067 | 15.79 | 1800 | 0.0267 | 0.7443 |
| 0.006 | 17.54 | 2000 | 0.0293 | 0.7416 |
Framework versions
- Transformers 4.37.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.0
- Tokenizers 0.15.2