shanhy's picture
End of training
c141fa9 verified
metadata
license: mit
base_model: FacebookAI/xlm-roberta-base
tags:
  - generated_from_trainer
model-index:
  - name: xlm-roberta-base_lr2e-05_seed42_basic_original_esp-kin-eng_train
    results: []

xlm-roberta-base_lr2e-05_seed42_basic_original_esp-kin-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: 0.0284
  • Spearman Corr: 0.7240

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 1.63 200 0.0303 0.6450
0.0362 3.27 400 0.0274 0.6899
0.0204 4.9 600 0.0238 0.6971
0.0161 6.53 800 0.0278 0.7107
0.0124 8.16 1000 0.0294 0.7245
0.0124 9.8 1200 0.0284 0.7332
0.0093 11.43 1400 0.0294 0.7160
0.0075 13.06 1600 0.0275 0.7203
0.0061 14.69 1800 0.0267 0.7250
0.0053 16.33 2000 0.0308 0.7266
0.0053 17.96 2200 0.0316 0.7213
0.0049 19.59 2400 0.0293 0.7216
0.0042 21.22 2600 0.0284 0.7240

Framework versions

  • Transformers 4.37.2
  • Pytorch 2.1.0+cu121
  • Datasets 2.17.0
  • Tokenizers 0.15.2