---
license: mit
tags:
- generated_from_trainer
datasets:
- multi_news
metrics:
- rouge
model-index:
- name: bart_large_summarise_v3
  results:
  - task:
      name: Sequence-to-sequence Language Modeling
      type: text2text-generation
    dataset:
      name: multi_news
      type: multi_news
      config: default
      split: train
      args: default
    metrics:
    - name: Rouge1
      type: rouge
      value: 0.3914
---

![SGH logo.png](https://s3.amazonaws.com/moonup/production/uploads/1667143139655-631feef1124782a19eff4243.png)

This model is a fine-tuned version of [facebook/bart-large-cnn](https://huggingface.co/facebook/bart-large-cnn) on the multi_news dataset.
It achieves the following results on the evaluation set:
- Loss: 4.1359
- Rouge1: 0.3914
- Rouge2: 0.1399
- Rougel: 0.2039
- Rougelsum: 0.3504
- Gen Len: 141.64

## Model description

This model was created to generate summaries of news articles.

## Intended uses & limitations

The model takes up to maximum article length of 1024 tokens and generates a summary of maximum length of 512 tokens.

## Training and evaluation data

This model was trained on 1000 articles and summaries from the Multi-News dataset. https://arxiv.org/abs/1906.01749

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 10
- label_smoothing_factor: 0.1

### Framework versions

- Transformers 4.23.1
- Pytorch 1.12.1+cu113
- Datasets 2.6.1
- Tokenizers 0.13.1