multi-news-model / README.md
cs608's picture
update model card README.md
66df073
metadata
license: apache-2.0
tags:
  - summarization
  - generated_from_trainer
datasets:
  - multi_news
metrics:
  - rouge
model-index:
  - name: CS685-text-summarizer
    results:
      - task:
          name: Sequence-to-sequence Language Modeling
          type: text2text-generation
        dataset:
          name: multi_news
          type: multi_news
          config: default
          split: train[:20%]
          args: default
        metrics:
          - name: Rouge1
            type: rouge
            value: 9.8777

CS685-text-summarizer

This model is a fine-tuned version of facebook/bart-base on the multi_news dataset. It achieves the following results on the evaluation set:

  • Loss: 2.4960
  • Rouge1: 9.8777
  • Rouge2: 3.6509
  • Rougel: 7.597
  • Rougelsum: 9.0495

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: 5.6e-05
  • train_batch_size: 6
  • eval_batch_size: 6
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 5

Training results

Training Loss Epoch Step Validation Loss Rouge1 Rouge2 Rougel Rougelsum
2.8402 1.0 1349 2.5256 9.7548 3.5429 7.5078 8.9174
2.5581 2.0 2698 2.5011 9.7256 3.6262 7.5644 8.8904
2.3908 3.0 4047 2.4948 9.7569 3.5976 7.5209 8.8919
2.2744 4.0 5396 2.4875 9.9116 3.7758 7.6823 9.0881
2.1908 5.0 6745 2.4960 9.8777 3.6509 7.597 9.0495

Framework versions

  • Transformers 4.29.1
  • Pytorch 2.0.0+cu118
  • Datasets 2.12.0
  • Tokenizers 0.13.3