README.md

metadata

license: apache-2.0
tags:
  - summarization
datasets:
  - multi_news
metrics:
  - rouge
model-index:
  - name: distilbart-cnn-12-6-ftn-multi_news
    results:
      - task:
          name: Sequence-to-sequence Language Modeling
          type: summarization
        dataset:
          name: multi_news
          type: multi_news
          args: default
        metrics:
          - name: Rouge1
            type: rouge
            value: 41.6136
      - task:
          type: summarization
          name: Summarization
        dataset:
          name: multi_news
          type: multi_news
          config: default
          split: test
        metrics:
          - name: ROUGE-1
            type: rouge
            value: 39.6512
            verified: true
          - name: ROUGE-2
            type: rouge
            value: 14.333
            verified: true
          - name: ROUGE-L
            type: rouge
            value: 21.5797
            verified: true
          - name: ROUGE-LSUM
            type: rouge
            value: 35.5793
            verified: true
          - name: loss
            type: loss
            value: 5.507579803466797
            verified: true
          - name: gen_len
            type: gen_len
            value: 132.1745
            verified: true

distilbart-cnn-12-6-ftn-multi_news

This model is a fine-tuned version of sshleifer/distilbart-cnn-12-6 on the multi_news dataset. It achieves the following results on the evaluation set:

• Loss: 3.8143
• Rouge1: 41.6136
• Rouge2: 14.7454
• Rougel: 23.3597
• Rougelsum: 36.1973
• Gen Len: 130.874

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
• lr_scheduler_warmup_steps: 500
• num_epochs: 1
• mixed_precision_training: Native AMP
• label_smoothing_factor: 0.1

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
3.8821	0.89	2000	3.8143	41.6136	14.7454	23.3597	36.1973	130.874

Framework versions

• Transformers 4.20.1
• Pytorch 1.11.0+cu113
• Datasets 2.3.2
• Tokenizers 0.12.1