metadata
license: apache-2.0
tags:
- summarization
- generated_from_trainer
datasets:
- billsum
metrics:
- rouge
model-index:
- name: billsum-full-data
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: billsum
type: billsum
config: default
split: train[:95%]
args: default
metrics:
- name: Rouge1
type: rouge
value: 18.0383
billsum-full-data
This model is a fine-tuned version of facebook/bart-base on the billsum dataset. It achieves the following results on the evaluation set:
- Loss: 1.6583
- Rouge1: 18.0383
- Rouge2: 14.8462
- Rougel: 17.6086
- Rougelsum: 17.6843
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: 2
- eval_batch_size: 2
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
|---|---|---|---|---|---|---|---|
| 2.1401 | 1.0 | 8101 | 1.8087 | 17.8461 | 14.6015 | 17.3956 | 17.4842 |
| 1.7596 | 2.0 | 16202 | 1.6980 | 18.0568 | 14.7833 | 17.6068 | 17.6999 |
| 1.5789 | 3.0 | 24303 | 1.6583 | 18.0383 | 14.8462 | 17.6086 | 17.6843 |
Framework versions
- Transformers 4.29.1
- Pytorch 2.0.0
- Datasets 2.12.0
- Tokenizers 0.13.3