<!-- 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-cnn-pubmed-arxiv-pubmed-arxiv-arxiv-v3-e12
This model is a fine-tuned version of theojolliffe/bart-cnn-pubmed-arxiv-pubmed-arxiv-arxiv on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.8501
- Rouge1: 56.1453
- Rouge2: 40.018
- Rougel: 43.5586
- Rougelsum: 54.4271
- Gen Len: 142.0
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: 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: 12
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
---|---|---|---|---|---|---|---|---|
No log | 1.0 | 398 | 0.8670 | 54.4613 | 34.7958 | 36.5841 | 51.9208 | 142.0 |
0.8276 | 2.0 | 796 | 0.8061 | 53.5804 | 34.5801 | 37.4643 | 51.1494 | 142.0 |
0.5318 | 3.0 | 1194 | 0.8146 | 53.7541 | 34.2446 | 37.5488 | 51.2475 | 142.0 |
0.3541 | 4.0 | 1592 | 0.7578 | 53.7645 | 34.874 | 38.3958 | 51.3075 | 142.0 |
0.3541 | 5.0 | 1990 | 0.7778 | 55.2787 | 37.5539 | 40.5489 | 52.8514 | 142.0 |
0.2386 | 6.0 | 2388 | 0.7810 | 55.2487 | 38.6522 | 41.466 | 53.379 | 142.0 |
0.1652 | 7.0 | 2786 | 0.7905 | 54.3618 | 37.4987 | 40.7348 | 52.2938 | 142.0 |
0.1152 | 8.0 | 3184 | 0.7934 | 54.4888 | 37.649 | 40.3582 | 52.3451 | 142.0 |
0.0942 | 9.0 | 3582 | 0.8220 | 55.5489 | 39.8493 | 42.2318 | 53.727 | 142.0 |
0.0942 | 10.0 | 3980 | 0.8331 | 55.7509 | 39.9491 | 43.2336 | 53.9748 | 142.0 |
0.0669 | 11.0 | 4378 | 0.8298 | 57.3881 | 42.6588 | 45.4694 | 55.8334 | 142.0 |
0.0531 | 12.0 | 4776 | 0.8501 | 56.1453 | 40.018 | 43.5586 | 54.4271 | 142.0 |
Framework versions
- Transformers 4.19.2
- Pytorch 1.11.0+cu113
- Datasets 2.2.2
- Tokenizers 0.12.1