Edit model card

recipe-roberta-is

This model is a fine-tuned version of roberta-base on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.8382

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: 256
  • eval_batch_size: 256
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 20
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
1.334 1.0 961 1.1217
1.1638 2.0 1922 1.0369
1.0936 3.0 2883 0.9922
1.0503 4.0 3844 0.9606
1.0188 5.0 4805 0.9314
0.9953 6.0 5766 0.9256
0.9769 7.0 6727 0.9109
0.9599 8.0 7688 0.8978
0.9461 9.0 8649 0.8813
0.9377 10.0 9610 0.8777
0.9253 11.0 10571 0.8755
0.918 12.0 11532 0.8601
0.9112 13.0 12493 0.8541
0.9043 14.0 13454 0.8548
0.8984 15.0 14415 0.8470
0.8958 16.0 15376 0.8412
0.8914 17.0 16337 0.8345
0.8882 18.0 17298 0.8353
0.8871 19.0 18259 0.8344
0.8839 20.0 19220 0.8382

Framework versions

  • Transformers 4.19.0.dev0
  • Pytorch 1.11.0+cu102
  • Datasets 2.3.2
  • Tokenizers 0.12.1
Downloads last month
13