wikinews_model
This model is a fine-tuned version of thanathorn/mt5-cpe-kmutt-thai-sentence-sum on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 4.1099
- Rouge1: 0.3228
- Rouge2: 0.1579
- Rougel: 0.3281
- Rougelsum: 0.3263
- Gen Len: 15.0789
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: 8
- eval_batch_size: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5000
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|---|---|---|---|---|---|---|---|---|
| 0.9441 | 35.71 | 500 | 2.3363 | 0.2632 | 0.1404 | 0.2632 | 0.2632 | 15.7895 |
| 0.2861 | 71.43 | 1000 | 3.0245 | 0.2632 | 0.1404 | 0.2632 | 0.2632 | 16.2105 |
| 0.0995 | 107.14 | 1500 | 3.5794 | 0.2895 | 0.1404 | 0.2895 | 0.2868 | 16.5 |
| 0.0489 | 142.86 | 2000 | 4.1099 | 0.3228 | 0.1579 | 0.3281 | 0.3263 | 15.0789 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1+cu116
- Datasets 2.10.1
- Tokenizers 0.13.2