bart-abs-1509-0313-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: 3.7689
• Rouge/rouge1: 0.4647
• Rouge/rouge2: 0.2065
• Rouge/rougel: 0.3953
• Rouge/rougelsum: 0.3967
• Bertscore/bertscore-precision: 0.8961
• Bertscore/bertscore-recall: 0.8941
• Bertscore/bertscore-f1: 0.895
• Meteor: 0.4195
• Gen Len: 37.9545

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
0.332	1.0	109	2.9323	0.4383	0.1841	0.3691	0.3701	0.8915	0.8887	0.8899	0.3951	37.2182
0.3384	2.0	218	3.0419	0.4611	0.2038	0.3901	0.3918	0.8941	0.8913	0.8925	0.4163	37.1909
0.2354	3.0	327	3.2793	0.445	0.1903	0.3776	0.3785	0.8938	0.8895	0.8915	0.394	36.4545
0.1736	4.0	436	3.4093	0.4545	0.2	0.3877	0.3885	0.8939	0.8921	0.8928	0.4094	38.3818
0.1406	5.0	545	3.5183	0.4634	0.2065	0.394	0.3945	0.898	0.8925	0.8951	0.4032	35.5182
0.1108	6.0	654	3.6131	0.4667	0.2075	0.3961	0.3964	0.8966	0.8936	0.8949	0.4155	37.4364
0.0906	7.0	763	3.6935	0.4602	0.2002	0.3892	0.3903	0.8922	0.8944	0.8931	0.4116	39.6
0.0788	8.0	872	3.7280	0.4704	0.212	0.4	0.4018	0.8941	0.8955	0.8946	0.431	39.8727
0.0708	9.0	981	3.7601	0.468	0.2062	0.3979	0.3994	0.8953	0.8948	0.8949	0.4244	38.9727
0.0637	10.0	1090	3.7689	0.4647	0.2065	0.3953	0.3967	0.8961	0.8941	0.895	0.4195	37.9545

Framework versions

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