bart-large-xsum_readme_summarization

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

• Loss: 2.1218
• Rouge1: 0.5637
• Rouge2: 0.4319
• Rougel: 0.5369
• Rougelsum: 0.5371
• Gen Len: 21.5048

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: 2e-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: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
2.1078	1.0	1458	1.9876	0.4994	0.3426	0.4684	0.4682	20.1103
1.3919	2.0	2916	1.8539	0.5137	0.3697	0.4841	0.4839	21.8345
0.9878	3.0	4374	1.9027	0.5441	0.401	0.5174	0.5171	20.1487
0.6594	4.0	5832	2.0362	0.5628	0.4272	0.5385	0.538	21.3417
0.4691	5.0	7290	2.1218	0.5637	0.4319	0.5369	0.5371	21.5048

Framework versions

• Transformers 4.35.1
• Pytorch 2.1.0+cu121
• Datasets 2.14.6
• Tokenizers 0.14.1