metadata
license: mit
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: bart-base-cnn-swe
results: []
bart-base-cnn-swe
This model is a fine-tuned version of Gabriel/bart-base-cnn-swe on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.9598
- Rouge1: 22.1847
- Rouge2: 10.4718
- Rougel: 18.2296
- Rougelsum: 20.8603
- Gen Len: 19.9964
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: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|---|---|---|---|---|---|---|---|---|
| 2.0969 | 1.0 | 17944 | 2.0137 | 22.0653 | 10.2668 | 18.0992 | 20.7207 | 19.9984 |
| 1.904 | 2.0 | 35888 | 1.9598 | 22.1847 | 10.4718 | 18.2296 | 20.8603 | 19.9964 |
Framework versions
- Transformers 4.22.1
- Pytorch 1.12.1+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1