metadata
license: mit
base_model: facebook/bart-large-cnn
tags:
- summarization
- generated_from_trainer
datasets:
- tldr_news
metrics:
- rouge
model-index:
- name: my_summ
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: tldr_news
type: tldr_news
config: all
split: test
args: all
metrics:
- name: Rouge1
type: rouge
value: 0.21647643221587914
my_summ
This model is a fine-tuned version of facebook/bart-large-cnn on the tldr_news dataset. It achieves the following results on the evaluation set:
- Loss: 4.1133
- Rouge1: 0.2165
- Rouge2: 0.0872
- Rougel: 0.1846
- Rougelsum: 0.1881
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: 5.6e-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: 8
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
|---|---|---|---|---|---|---|---|
| 2.2607 | 1.0 | 125 | 2.2706 | 0.2318 | 0.0950 | 0.1983 | 0.2024 |
| 1.1698 | 2.0 | 250 | 2.3624 | 0.2150 | 0.0848 | 0.1828 | 0.1856 |
| 0.5798 | 3.0 | 375 | 2.8369 | 0.2144 | 0.0838 | 0.1802 | 0.1848 |
| 0.2813 | 4.0 | 500 | 3.3045 | 0.2112 | 0.0803 | 0.1788 | 0.1821 |
| 0.1544 | 5.0 | 625 | 3.6092 | 0.2096 | 0.0793 | 0.1780 | 0.1838 |
| 0.0862 | 6.0 | 750 | 3.7615 | 0.2168 | 0.0848 | 0.1851 | 0.1881 |
| 0.0518 | 7.0 | 875 | 3.9039 | 0.2180 | 0.0861 | 0.1842 | 0.1873 |
| 0.0253 | 8.0 | 1000 | 4.1133 | 0.2165 | 0.0872 | 0.1846 | 0.1881 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.0