encoder_decoder_es

This model is a fine-tuned version of on the cc_news_es_titles dataset. It achieves the following results on the evaluation set:

  • Loss: 7.8773
  • Rouge2 Precision: 0.002
  • Rouge2 Recall: 0.0116
  • Rouge2 Fmeasure: 0.0034

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: 0.003
  • train_batch_size: 32
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • num_epochs: 4
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Rouge2 Precision Rouge2 Recall Rouge2 Fmeasure
7.8807 1.0 5784 7.8976 0.0023 0.012 0.0038
7.8771 2.0 11568 7.8873 0.0018 0.0099 0.003
7.8588 3.0 17352 7.8819 0.0015 0.0085 0.0025
7.8507 4.0 23136 7.8773 0.002 0.0116 0.0034

Framework versions

  • Transformers 4.12.3
  • Pytorch 1.9.1
  • Datasets 1.15.1
  • Tokenizers 0.10.3
Downloads last month
3
Hosted inference API
This model can be loaded on the Inference API on-demand.

Evaluation results