bart-abs-2409-1947-lr-3e-05-bs-8-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: 4.2084
• Rouge/rouge1: 0.4731
• Rouge/rouge2: 0.2204
• Rouge/rougel: 0.4091
• Rouge/rougelsum: 0.4105
• Bertscore/bertscore-precision: 0.8963
• Bertscore/bertscore-recall: 0.8954
• Bertscore/bertscore-f1: 0.8957
• Meteor: 0.4281
• Gen Len: 38.3182

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: 8
• eval_batch_size: 8
• 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
0.0785	1.0	109	3.9838	0.4581	0.1987	0.3887	0.3892	0.8957	0.8906	0.893	0.4019	35.6455
0.0862	2.0	218	3.9075	0.4574	0.2028	0.3856	0.3863	0.8923	0.8918	0.8919	0.4091	39.1545
0.0792	3.0	327	3.9794	0.4593	0.1972	0.3807	0.3814	0.889	0.8926	0.8906	0.4145	41.9182
0.0673	4.0	436	4.0419	0.4634	0.2047	0.3928	0.3937	0.8948	0.8918	0.8931	0.4133	36.6909
0.0604	5.0	545	4.1048	0.4629	0.2112	0.396	0.3971	0.8956	0.8926	0.8939	0.4118	36.9727
0.0548	6.0	654	4.1331	0.4556	0.2042	0.3904	0.391	0.8938	0.8917	0.8926	0.4079	38.1545
0.0508	7.0	763	4.1740	0.4546	0.1949	0.383	0.3842	0.8925	0.8903	0.8913	0.4028	37.5273
0.0473	8.0	872	4.1643	0.4653	0.212	0.401	0.4026	0.8949	0.8939	0.8942	0.4212	38.4818
0.0438	9.0	981	4.1913	0.472	0.2155	0.4063	0.4071	0.8969	0.8947	0.8956	0.4223	37.9091
0.0401	10.0	1090	4.2084	0.4731	0.2204	0.4091	0.4105	0.8963	0.8954	0.8957	0.4281	38.3182

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0
• Datasets 3.0.0
• Tokenizers 0.19.1