metadata
license: mit
base_model: facebook/bart-large-cnn
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: BART-CNN-Convosumm
results: []
BART-CNN-Convosumm
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: 3.8797
- Rouge1: 38.6252
- Rouge2: 12.2556
- Rougel: 23.902
- Rougelsum: 34.6324
- Gen Len: 81.28
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: 1
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 20
- total_train_batch_size: 20
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: polynomial
- lr_scheduler_warmup_steps: 1
- num_epochs: 7
- label_smoothing_factor: 0.1
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|---|---|---|---|---|---|---|---|---|
| 6.207 | 1.0 | 10 | 4.2651 | 32.3341 | 7.812 | 20.0411 | 29.4849 | 77.38 |
| 4.0248 | 1.99 | 20 | 3.9903 | 36.0787 | 11.0447 | 21.3596 | 33.2903 | 130.58 |
| 3.5933 | 2.99 | 30 | 3.9020 | 34.2931 | 11.2036 | 20.7935 | 30.8361 | 140.02 |
| 3.3086 | 3.98 | 40 | 3.8712 | 38.4842 | 11.9947 | 23.4913 | 34.4347 | 85.78 |
| 3.112 | 4.98 | 50 | 3.8700 | 38.652 | 11.8315 | 23.5208 | 34.5998 | 76.2 |
| 2.9933 | 5.97 | 60 | 3.8809 | 38.66 | 12.3337 | 23.4394 | 35.1976 | 83.26 |
| 2.834 | 6.97 | 70 | 3.8797 | 38.6252 | 12.2556 | 23.902 | 34.6324 | 81.28 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.0.0
- Datasets 2.1.0
- Tokenizers 0.15.0