File size: 2,025 Bytes
dc4a8e1 |
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-cnn
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: bart-large-xsumfinetuned-samsum
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-xsumfinetuned-samsum
This model is a fine-tuned version of [facebook/bart-large-cnn](https://huggingface.co/facebook/bart-large-cnn) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 2.9976
- Rouge1: 0.4246
- Rouge2: 0.2069
- Rougel: 0.3253
- Rougelsum: 0.3907
## 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: 6
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
|:-------------:|:-----:|:-----:|:---------------:|:------:|:------:|:------:|:---------:|
| 0.2858 | 1.0 | 3683 | 2.3127 | 0.4051 | 0.1926 | 0.3072 | 0.3720 |
| 0.3128 | 2.0 | 7366 | 2.3467 | 0.4007 | 0.1911 | 0.3037 | 0.3687 |
| 0.2544 | 3.0 | 11049 | 2.3126 | 0.4145 | 0.2019 | 0.3159 | 0.3801 |
| 0.1846 | 4.0 | 14732 | 2.6484 | 0.4088 | 0.1977 | 0.3096 | 0.3774 |
| 0.1143 | 5.0 | 18415 | 2.7793 | 0.4173 | 0.1997 | 0.3155 | 0.3843 |
| 0.0687 | 6.0 | 22098 | 2.9976 | 0.4246 | 0.2069 | 0.3253 | 0.3907 |
### Framework versions
- Transformers 4.40.1
- Pytorch 2.2.1+cu121
- Datasets 2.19.0
- Tokenizers 0.19.1
|