bart-abs-2409-0144-lr-3e-05-bs-4-maxep-10

This model is a fine-tuned version of sshleifer/distilbart-xsum-12-6 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.7398
• Rouge/rouge1: 0.4519
• Rouge/rouge2: 0.1957
• Rouge/rougel: 0.3795
• Rouge/rougelsum: 0.3809
• Bertscore/bertscore-precision: 0.8939
• Bertscore/bertscore-recall: 0.8913
• Bertscore/bertscore-f1: 0.8925
• Meteor: 0.4012
• Gen Len: 37.1455

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: 3e-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: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge/rouge1	Rouge/rouge2	Rouge/rougel	Rouge/rougelsum	Bertscore/bertscore-precision	Bertscore/bertscore-recall	Bertscore/bertscore-f1	Meteor	Gen Len
0.7649	1.0	217	2.3871	0.4493	0.2047	0.3867	0.3879	0.8969	0.89	0.8933	0.3916	35.6364
0.5599	2.0	434	2.5775	0.4553	0.2005	0.3849	0.3861	0.8951	0.8914	0.8931	0.3951	36.1636
0.4078	3.0	651	2.9176	0.4622	0.2118	0.3904	0.3927	0.8942	0.8925	0.8932	0.4137	36.9818
0.2969	4.0	868	3.1512	0.4589	0.2038	0.3877	0.3892	0.8957	0.8886	0.892	0.3961	34.6364
0.2291	5.0	1085	3.3475	0.4594	0.2035	0.3899	0.3915	0.8964	0.8925	0.8943	0.4099	36.9364
0.2006	6.0	1302	3.3661	0.466	0.209	0.3934	0.3959	0.896	0.8933	0.8945	0.4136	37.2818
0.1485	7.0	1519	3.5165	0.4639	0.2054	0.3846	0.3862	0.8939	0.8931	0.8934	0.4137	38.7636
0.1131	8.0	1736	3.6478	0.4595	0.202	0.3882	0.3908	0.8958	0.8903	0.8929	0.402	35.2727
0.0945	9.0	1953	3.7024	0.4614	0.2048	0.39	0.391	0.8933	0.894	0.8935	0.4163	40.1545
0.0794	10.0	2170	3.7398	0.4519	0.1957	0.3795	0.3809	0.8939	0.8913	0.8925	0.4012	37.1455

Framework versions

• Transformers 4.44.0
• Pytorch 2.4.0
• Datasets 2.21.0
• Tokenizers 0.19.1