metadata
license: apache-2.0
base_model: facebook/bart-base
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: bart_samsum_model
results: []
bart_samsum_model
This model is a fine-tuned version of facebook/bart-base on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.5246
- Rouge1: 0.472
- Rouge2: 0.2373
- Rougel: 0.3986
- Rougelsum: 0.3987
- Gen Len: 18.1954
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: 2.8410039143956672e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|---|---|---|---|---|---|---|---|---|
| 1.822 | 1.0 | 1842 | 1.5989 | 0.4581 | 0.2227 | 0.3846 | 0.385 | 18.1893 |
| 1.5641 | 2.0 | 3684 | 1.5492 | 0.4661 | 0.2328 | 0.3936 | 0.3936 | 18.2576 |
| 1.4168 | 3.0 | 5526 | 1.5246 | 0.472 | 0.2373 | 0.3986 | 0.3987 | 18.1954 |
Framework versions
- Transformers 4.37.2
- Pytorch 2.1.0+cu121
- Datasets 2.17.0
- Tokenizers 0.15.1