metadata
license: mit
tags:
- qmsum-summarization
- generated_from_trainer
metrics:
- rouge
model-index:
- name: bart-large-cnn-finetuned-qmsum-2-4
results: []
bart-large-cnn-finetuned-qmsum-2-4
This model is a fine-tuned version of facebook/bart-large-cnn on the None dataset. It achieves the following results on the evaluation set:
- Loss: 4.0277
- Rouge1: 0.3053
- Rouge2: 0.0660
- Rougel: 0.1903
- Rougelsum: 0.2598
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: 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: 4
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
|---|---|---|---|---|---|---|---|
| 3.3773 | 1.0 | 629 | 3.2522 | 0.2964 | 0.0713 | 0.1958 | 0.2593 |
| 2.3656 | 2.0 | 1258 | 3.2001 | 0.2942 | 0.0694 | 0.1921 | 0.2540 |
| 1.5843 | 3.0 | 1887 | 3.4248 | 0.3086 | 0.0687 | 0.1938 | 0.2648 |
| 0.9854 | 4.0 | 2516 | 4.0277 | 0.3053 | 0.0660 | 0.1903 | 0.2598 |
Framework versions
- Transformers 4.21.3
- Pytorch 1.12.1+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1