File size: 2,027 Bytes
5799831 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 |
---
license: mit
base_model: facebook/bart-large-xsum
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: bart-large-xsum_readme_summarization
results: []
---
<!-- 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-large-xsum_readme_summarization
This model is a fine-tuned version of [facebook/bart-large-xsum](https://huggingface.co/facebook/bart-large-xsum) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.1218
- Rouge1: 0.5637
- Rouge2: 0.4319
- Rougel: 0.5369
- Rougelsum: 0.5371
- Gen Len: 21.5048
## 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: 4
- eval_batch_size: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:------:|:------:|:------:|:---------:|:-------:|
| 2.1078 | 1.0 | 1458 | 1.9876 | 0.4994 | 0.3426 | 0.4684 | 0.4682 | 20.1103 |
| 1.3919 | 2.0 | 2916 | 1.8539 | 0.5137 | 0.3697 | 0.4841 | 0.4839 | 21.8345 |
| 0.9878 | 3.0 | 4374 | 1.9027 | 0.5441 | 0.401 | 0.5174 | 0.5171 | 20.1487 |
| 0.6594 | 4.0 | 5832 | 2.0362 | 0.5628 | 0.4272 | 0.5385 | 0.538 | 21.3417 |
| 0.4691 | 5.0 | 7290 | 2.1218 | 0.5637 | 0.4319 | 0.5369 | 0.5371 | 21.5048 |
### Framework versions
- Transformers 4.35.1
- Pytorch 2.1.0+cu121
- Datasets 2.14.6
- Tokenizers 0.14.1
|