metadata
license: mit
tags:
- generated_from_trainer
inference:
parameters:
temperature: 0.7
min_length: 30
max_length: 120
num_beams: 5
metrics:
- rouge
model-index:
- name: bart-base-cnn-xsum-swe
results: []
bart-base-cnn-xsum-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: 2.1895
- Rouge1: 31.1693
- Rouge2: 12.7388
- Rougel: 25.7655
- Rougelsum: 25.7862
- Gen Len: 19.7733
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: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 8
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|---|---|---|---|---|---|---|---|---|
| 2.3079 | 1.0 | 6375 | 2.1998 | 29.7845 | 11.125 | 24.3181 | 24.3562 | 19.7119 |
| 2.064 | 2.0 | 12750 | 2.1245 | 30.4641 | 11.7383 | 25.0254 | 25.0633 | 19.653 |
| 1.8647 | 3.0 | 19125 | 2.1005 | 30.8903 | 12.2265 | 25.3996 | 25.4252 | 19.7457 |
| 1.7098 | 4.0 | 25500 | 2.1073 | 31.1173 | 12.4124 | 25.6553 | 25.6913 | 19.7546 |
| 1.5761 | 5.0 | 31875 | 2.1227 | 30.9586 | 12.4907 | 25.5474 | 25.5745 | 19.7675 |
| 1.4618 | 6.0 | 38250 | 2.1484 | 31.115 | 12.6546 | 25.684 | 25.7151 | 19.7456 |
| 1.3643 | 7.0 | 44625 | 2.1705 | 31.2225 | 12.8069 | 25.7901 | 25.8154 | 19.7842 |
| 1.2944 | 8.0 | 51000 | 2.1895 | 31.1693 | 12.7388 | 25.7655 | 25.7862 | 19.7733 |
Framework versions
- Transformers 4.22.1
- Pytorch 1.12.1+cu113
- Datasets 2.5.1
- Tokenizers 0.12.1