metadata
license: mit
base_model: FacebookAI/xlm-roberta-base
tags:
- generated_from_trainer
model-index:
- name: xlm-roberta-base_lr2e-05_seed42_basic_original_kin-amh-eng_train
results: []
xlm-roberta-base_lr2e-05_seed42_basic_original_kin-amh-eng_train
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: 0.0275
- Spearman Corr: 0.7451
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 | 0.0329 | 0.6825 |
0.0465 | 3.51 | 400 | 0.0228 | 0.7144 |
0.0206 | 5.26 | 600 | 0.0245 | 0.7461 |
0.0149 | 7.02 | 800 | 0.0302 | 0.7391 |
0.0117 | 8.77 | 1000 | 0.0246 | 0.7478 |
0.0093 | 10.53 | 1200 | 0.0271 | 0.7429 |
0.008 | 12.28 | 1400 | 0.0288 | 0.7515 |
0.0067 | 14.04 | 1600 | 0.0270 | 0.7478 |
0.0067 | 15.79 | 1800 | 0.0270 | 0.7492 |
0.0062 | 17.54 | 2000 | 0.0274 | 0.7397 |
0.0051 | 19.3 | 2200 | 0.0258 | 0.7481 |
0.0047 | 21.05 | 2400 | 0.0275 | 0.7451 |
Framework versions
- Transformers 4.37.2
- Pytorch 2.1.0+cu121
- Datasets 2.17.0
- Tokenizers 0.15.2