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