bart-samsum-finetuned

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

• Loss: 0.1326

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: 1e-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: 100
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
0.1196	1.0	74	0.1362
0.0948	2.0	148	0.1334
0.0738	3.0	222	0.1326

Evaluation results

Rouge Scores:

Metric	Precision	Recall	F-Measure
rouge1	Low - 0.2923	Low - 0.5755	Low - 0.3645
	Mid - 0.3012	Mid - 0.5881	Mid - 0.3722
	High - 0.3108	High - 0.6011	High - 0.3811
rouge2	Low - 0.1185	Low - 0.2418	Low - 0.1481
	Mid - 0.1252	Mid - 0.2545	Mid - 0.1555
	High - 0.1321	High - 0.2682	High - 0.1632
rougeL	Low - 0.2182	Low - 0.4434	Low - 0.2744
	Mid - 0.2251	Mid - 0.4547	Mid - 0.2810
	High - 0.2328	High - 0.4679	High - 0.2886
rougeLsum	Low - 0.2178	Low - 0.4425	Low - 0.2739
	Mid - 0.2249	Mid - 0.4546	Mid - 0.2807
	High - 0.2321	High - 0.4679	High - 0.2883

BERTScore:

Precision	Recall	F1
0.6054495	0.6918860	0.6425597

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2