kobart-base-v2-finetuned-paper
This model is a fine-tuned version of gogamza/kobart-base-v2 on the aihub_paper_summarization dataset. It achieves the following results on the evaluation set:
- Loss: 1.2966
- Rouge1: 6.2883
- Rouge2: 1.7038
- Rougel: 6.2556
- Rougelsum: 6.2618
- Gen Len: 20.0
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: 2e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|---|---|---|---|---|---|---|---|---|
| 1.2215 | 1.0 | 8831 | 1.3293 | 6.2425 | 1.7317 | 6.2246 | 6.2247 | 20.0 |
| 1.122 | 2.0 | 17662 | 1.3056 | 6.2298 | 1.7005 | 6.2042 | 6.2109 | 20.0 |
| 1.0914 | 3.0 | 26493 | 1.2966 | 6.2883 | 1.7038 | 6.2556 | 6.2618 | 20.0 |
Framework versions
- Transformers 4.25.1
- Pytorch 1.13.1
- Datasets 2.8.0
- Tokenizers 0.13.2