Edit model card

PegasusXSUM_GNAD

This model is a fine-tuned version of Einmalumdiewelt/PegasusXSUM_GNAD on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 2.4386
  • Rouge1: 26.7818
  • Rouge2: 7.6864
  • Rougel: 18.6264
  • Rougelsum: 22.822
  • Gen Len: 67.076

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: 5e-05
  • train_batch_size: 2
  • eval_batch_size: 2
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 10.0

Training results

Framework versions

  • Transformers 4.22.0.dev0
  • Pytorch 1.12.0+cu113
  • Datasets 2.4.0
  • Tokenizers 0.12.1
Downloads last month
55

Space using Einmalumdiewelt/PegasusXSUM_GNAD 1