metadata
license: apache-2.0
base_model: sshleifer/distilbart-xsum-12-6
tags:
- generated_from_trainer
model-index:
- name: bart-abs-1409-1800-lr-3e-05-bs-4-maxep-6
results: []
bart-abs-1409-1800-lr-3e-05-bs-4-maxep-6
This model is a fine-tuned version of sshleifer/distilbart-xsum-12-6 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 3.3829
- Rouge/rouge1: 0.3549
- Rouge/rouge2: 0.1363
- Rouge/rougel: 0.3021
- Rouge/rougelsum: 0.3032
- Bertscore/bertscore-precision: 0.9037
- Bertscore/bertscore-recall: 0.8687
- Bertscore/bertscore-f1: 0.8857
- Meteor: 0.2545
- Gen Len: 25.5
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: 3e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 6
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge/rouge1 | Rouge/rouge2 | Rouge/rougel | Rouge/rougelsum | Bertscore/bertscore-precision | Bertscore/bertscore-recall | Bertscore/bertscore-f1 | Meteor | Gen Len |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1.7439 | 1.0 | 13 | 2.8127 | 0.2992 | 0.0924 | 0.2491 | 0.2492 | 0.8984 | 0.8552 | 0.876 | 0.2093 | 22.5 |
| 1.263 | 2.0 | 26 | 2.9358 | 0.3773 | 0.1311 | 0.3058 | 0.306 | 0.9115 | 0.8688 | 0.8895 | 0.2469 | 24.9 |
| 0.8455 | 3.0 | 39 | 3.0548 | 0.4307 | 0.1554 | 0.3399 | 0.3393 | 0.9029 | 0.8741 | 0.8881 | 0.3255 | 28.1 |
| 0.6581 | 4.0 | 52 | 3.2153 | 0.3986 | 0.1569 | 0.3487 | 0.3471 | 0.9111 | 0.8757 | 0.8929 | 0.3092 | 25.8 |
| 0.4915 | 5.0 | 65 | 3.2936 | 0.4004 | 0.1256 | 0.3204 | 0.3222 | 0.9035 | 0.8738 | 0.8883 | 0.2892 | 26.0 |
| 0.393 | 6.0 | 78 | 3.3829 | 0.3549 | 0.1363 | 0.3021 | 0.3032 | 0.9037 | 0.8687 | 0.8857 | 0.2545 | 25.5 |
Framework versions
- Transformers 4.44.0
- Pytorch 2.4.0
- Datasets 2.21.0
- Tokenizers 0.19.1