bart-base-summarization-medical_on_cnn-51

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

• Loss: 3.3788
• Rouge1: 0.2533
• Rouge2: 0.0944
• Rougel: 0.2004
• Rougelsum: 0.2239
• Gen Len: 18.404

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: 1
• seed: 51
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 6
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
2.7137	1.0	1250	3.3745	0.2514	0.0905	0.1971	0.2211	19.121
2.5925	2.0	2500	3.3611	0.2522	0.0936	0.1993	0.2232	18.703
2.593	3.0	3750	3.3672	0.2543	0.094	0.2003	0.226	18.644
2.5582	4.0	5000	3.3758	0.2522	0.093	0.1993	0.2231	18.403
2.5285	5.0	6250	3.3766	0.2534	0.0943	0.2006	0.2243	18.416
2.5327	6.0	7500	3.3788	0.2533	0.0944	0.2004	0.2239	18.404

Framework versions

• PEFT 0.12.0
• Transformers 4.42.4
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1