File size: 5,439 Bytes
5e09fbc |
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 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 |
---
license: mit
base_model: facebook/bart-large-xsum
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: text_shortening_model_v37
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. -->
# text_shortening_model_v37
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: 2.9472
- Rouge1: 0.4923
- Rouge2: 0.2809
- Rougel: 0.4462
- Rougelsum: 0.4468
- Bert precision: 0.8731
- Bert recall: 0.8773
- Average word count: 9.1021
- Max word count: 15
- Min word count: 5
- Average token count: 16.8198
- % shortened texts with length > 12: 8.7087
## 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: 0.0003
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 15
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Bert precision | Bert recall | Average word count | Max word count | Min word count | Average token count | % shortened texts with length > 12 |
|:-------------:|:-----:|:----:|:---------------:|:------:|:------:|:------:|:---------:|:--------------:|:-----------:|:------------------:|:--------------:|:--------------:|:-------------------:|:----------------------------------:|
| 1.5911 | 1.0 | 73 | 1.8586 | 0.4823 | 0.2756 | 0.4416 | 0.4423 | 0.8661 | 0.8758 | 8.9399 | 21 | 4 | 16.9489 | 7.8078 |
| 0.9246 | 2.0 | 146 | 2.2274 | 0.4039 | 0.2049 | 0.3771 | 0.3764 | 0.8526 | 0.855 | 8.0991 | 13 | 4 | 14.6006 | 0.6006 |
| 0.7574 | 3.0 | 219 | 1.8752 | 0.4463 | 0.2263 | 0.4072 | 0.4071 | 0.8629 | 0.8654 | 8.3934 | 14 | 5 | 14.3303 | 3.003 |
| 0.6131 | 4.0 | 292 | 1.8338 | 0.4896 | 0.2691 | 0.4451 | 0.4456 | 0.8747 | 0.8711 | 7.982 | 13 | 4 | 13.9249 | 0.3003 |
| 0.4422 | 5.0 | 365 | 1.8257 | 0.492 | 0.2727 | 0.4499 | 0.4504 | 0.8734 | 0.875 | 8.5165 | 16 | 5 | 14.4595 | 3.003 |
| 0.4227 | 6.0 | 438 | 2.1249 | 0.4666 | 0.2475 | 0.418 | 0.4178 | 0.8657 | 0.8697 | 9.3874 | 16 | 4 | 16.9399 | 8.4084 |
| 0.3714 | 7.0 | 511 | 2.1010 | 0.4838 | 0.274 | 0.436 | 0.4364 | 0.869 | 0.8754 | 9.4264 | 16 | 5 | 14.9369 | 9.009 |
| 0.2638 | 8.0 | 584 | 2.0803 | 0.489 | 0.2799 | 0.4404 | 0.4404 | 0.8701 | 0.8751 | 8.976 | 15 | 4 | 15.5736 | 8.4084 |
| 0.2103 | 9.0 | 657 | 2.1093 | 0.4888 | 0.2722 | 0.4381 | 0.438 | 0.872 | 0.8751 | 9.1952 | 16 | 5 | 16.7447 | 9.9099 |
| 0.1475 | 10.0 | 730 | 2.3159 | 0.4684 | 0.2597 | 0.4243 | 0.4244 | 0.8632 | 0.8721 | 9.4234 | 15 | 5 | 16.8288 | 11.7117 |
| 0.122 | 11.0 | 803 | 2.4090 | 0.4845 | 0.2729 | 0.4421 | 0.4427 | 0.8721 | 0.8748 | 8.8018 | 16 | 5 | 16.4264 | 5.7057 |
| 0.0915 | 12.0 | 876 | 2.6598 | 0.4838 | 0.2691 | 0.4376 | 0.437 | 0.8698 | 0.8742 | 9.1652 | 16 | 5 | 16.9009 | 10.2102 |
| 0.073 | 13.0 | 949 | 2.5266 | 0.4973 | 0.2861 | 0.4479 | 0.4495 | 0.8743 | 0.8776 | 9.0631 | 16 | 5 | 16.5796 | 8.4084 |
| 0.0526 | 14.0 | 1022 | 2.7673 | 0.4955 | 0.2821 | 0.4464 | 0.4463 | 0.8716 | 0.8791 | 9.4685 | 16 | 5 | 17.2012 | 10.5105 |
| 0.042 | 15.0 | 1095 | 2.9472 | 0.4923 | 0.2809 | 0.4462 | 0.4468 | 0.8731 | 0.8773 | 9.1021 | 15 | 5 | 16.8198 | 8.7087 |
### Framework versions
- Transformers 4.33.1
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.13.3
|