metadata

language: sv
license: mit
metrics:
  - rouge
model-index:
  - name: bart-base-cnn-swe
    results:
      - task:
          type: summarization
          name: summarization
        dataset:
          name: Gabriel/cnn_daily_swe
          type: Gabriel/cnn_daily_swe
          split: validation
        metrics:
          - name: Validation ROGUE-1
            type: rouge-1
            value: 21.7291
            verified: true
          - name: Validation ROGUE-2
            type: rouge-2
            value: 10.0209
            verified: true
          - name: Validation ROGUE-L
            type: rouge-l
            value: 17.775
            verified: true
datasets:
  - Gabriel/cnn_daily_swe
tags:
  - summarization
widget: null
'-text': >-
  Minst två personer dödades i en misstänkt bomb attack på en personbuss i de
  stridiga södra filippiner på måndag, sade militären.

bart-base-cnn-swe

This model is a fine-tuned version of KBLab/bart-base-swedish-cased on the None dataset. It achieves the following results on the evaluation set:

Loss: 2.1656
Rouge1: 21.7291
Rouge2: 10.0209
Rougel: 17.775
Rougelsum: 20.429
Gen Len: 19.9931

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: 1
eval_batch_size: 1
seed: 42
gradient_accumulation_steps: 16
total_train_batch_size: 16
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 1
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
2.3512	1.0	17944	2.1656	21.7291	10.0209	17.775	20.429	19.9931

Framework versions

Transformers 4.21.2
Pytorch 1.12.1+cu113
Datasets 2.4.0
Tokenizers 0.12.1