bart-base-finetuned-multi-news

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

• Loss: 2.6353
• Rouge1: 15.1146
• Rouge2: 5.3873
• Rougel: 11.4132
• Rougelsum: 13.2739

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 7

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
2.9189	1.0	625	2.4645	15.2063	5.2852	11.5864	13.4208
2.4697	2.0	1250	2.4706	15.3737	5.4725	11.7465	13.5681
2.1831	3.0	1875	2.4789	14.8306	5.0857	11.2416	13.1072
1.9598	4.0	2500	2.5299	15.1744	5.5465	11.6445	13.4053
1.7777	5.0	3125	2.5799	14.9417	5.2124	11.3553	13.1401
1.6454	6.0	3750	2.6028	14.9804	5.333	11.294	13.2385
1.554	7.0	4375	2.6353	15.1146	5.3873	11.4132	13.2739

Framework versions

• Transformers 4.28.1
• Pytorch 2.0.0+cu118
• Datasets 2.11.0
• Tokenizers 0.13.3