bart-abs-1509-0313-lr-3e-05-bs-2-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: 4.0180
• Rouge/rouge1: 0.4724
• Rouge/rouge2: 0.2094
• Rouge/rougel: 0.3964
• Rouge/rougelsum: 0.3976
• Bertscore/bertscore-precision: 0.8964
• Bertscore/bertscore-recall: 0.8932
• Bertscore/bertscore-f1: 0.8947
• Meteor: 0.4217
• Gen Len: 36.8818

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: 2
• eval_batch_size: 2
• 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
0.6813	1.0	434	2.4588	0.4636	0.2101	0.3907	0.3921	0.8975	0.8904	0.8937	0.4099	35.0727
0.6702	2.0	868	2.5377	0.4448	0.1862	0.3725	0.3735	0.8942	0.8887	0.8913	0.3825	35.7273
0.4591	3.0	1302	2.8762	0.4533	0.1916	0.3767	0.3778	0.8961	0.8897	0.8928	0.3911	35.2091
0.3165	4.0	1736	3.2129	0.4519	0.1976	0.3803	0.3806	0.8936	0.891	0.8922	0.4023	37.6364
0.2222	5.0	2170	3.4971	0.47	0.2049	0.392	0.3924	0.8959	0.8926	0.8941	0.4107	36.5545
0.1596	6.0	2604	3.6405	0.4607	0.2101	0.3853	0.3879	0.8943	0.8908	0.8924	0.4021	37.2273
0.1166	7.0	3038	3.7827	0.4759	0.2191	0.4086	0.4106	0.8988	0.8928	0.8956	0.4173	35.5727
0.0891	8.0	3472	3.9388	0.4677	0.2047	0.3905	0.3925	0.8933	0.8927	0.8929	0.417	38.7
0.0695	9.0	3906	3.9583	0.4775	0.2116	0.4032	0.4051	0.8981	0.8931	0.8955	0.4228	36.3182
0.0592	10.0	4340	4.0180	0.4724	0.2094	0.3964	0.3976	0.8964	0.8932	0.8947	0.4217	36.8818

Framework versions

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