<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
bart-base-finetuned-summscreen-bestval-100-genlen-10-epochs
This model is a fine-tuned version of facebook/bart-base on the SummScreen dataset. It achieves the following results on the evaluation set:
- Loss: 3.0979
- Rouge1: 31.5373
- Rouge2: 6.6821
- Rougel: 18.6754
- Rougelsum: 27.4448
- Gen Len: 80.1927
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: 1e-05
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|---|---|---|---|---|---|---|---|---|
| 3.4849 | 0.99 | 3500 | 3.2071 | 28.6828 | 5.2634 | 17.218 | 25.487 | 94.059 |
| 3.2933 | 1.99 | 7000 | 3.1329 | 29.9774 | 5.7038 | 17.7705 | 26.2492 | 88.2358 |
| 3.1088 | 2.98 | 10500 | 3.1010 | 29.6903 | 5.6976 | 17.7468 | 25.9472 | 81.3129 |
| 2.9605 | 3.98 | 14000 | 3.0811 | 30.2088 | 6.1092 | 18.157 | 26.3051 | 77.8844 |
| 2.8778 | 4.97 | 17500 | 3.0747 | 30.6996 | 6.3038 | 18.4725 | 26.8669 | 81.6168 |
| 2.788 | 5.97 | 21000 | 3.0896 | 30.7478 | 6.4468 | 18.3755 | 26.8789 | 85.6395 |
| 2.7218 | 6.96 | 24500 | 3.0961 | 30.994 | 6.4407 | 18.4929 | 26.9802 | 79.1315 |
| 2.6753 | 7.96 | 28000 | 3.0892 | 31.336 | 6.6768 | 18.8122 | 27.389 | 83.2313 |
| 2.5753 | 8.95 | 31500 | 3.0960 | 31.3248 | 6.4093 | 18.6552 | 27.2087 | 80.1474 |
| 2.5918 | 9.95 | 35000 | 3.0979 | 31.5373 | 6.6821 | 18.6754 | 27.4448 | 80.1927 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.13.1
- Datasets 2.9.0
- Tokenizers 0.13.2
