File size: 1,923 Bytes
30def23 |
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 |
---
license: mit
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: bart-cnn-pubmed-arxiv-v3-e4
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-cnn-pubmed-arxiv-v3-e4
This model is a fine-tuned version of [theojolliffe/bart-cnn-pubmed-arxiv](https://huggingface.co/theojolliffe/bart-cnn-pubmed-arxiv) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.7934
- Rouge1: 54.2624
- Rouge2: 35.6024
- Rougel: 37.1697
- Rougelsum: 51.5144
- Gen Len: 141.9815
## 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: 4
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:--------:|
| No log | 1.0 | 398 | 0.9533 | 52.3191 | 32.4576 | 33.2016 | 49.6502 | 142.0 |
| 1.1154 | 2.0 | 796 | 0.8407 | 53.6639 | 34.3433 | 36.1893 | 50.9077 | 142.0 |
| 0.6856 | 3.0 | 1194 | 0.7978 | 54.4723 | 36.1315 | 37.7891 | 51.902 | 142.0 |
| 0.4943 | 4.0 | 1592 | 0.7934 | 54.2624 | 35.6024 | 37.1697 | 51.5144 | 141.9815 |
### Framework versions
- Transformers 4.18.0
- Pytorch 1.11.0+cu113
- Datasets 2.1.0
- Tokenizers 0.12.1
|