BMU_Finetuned_BART-large_MedQuad

This model is a fine-tuned version of facebook/bart-large on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.5144
• Rouge1: 36.45
• Rouge2: 19.67
• Rougel: 27.7
• Rougelsum: 34.52
• Meteor: 24.99
• Bertscore Precision: 82.36
• Bertscore Recall: 79.28
• Bertscore F1: 80.68
• Gen Len: 152.88

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: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Meteor	Bertscore Precision	Bertscore Recall	Bertscore F1	Gen Len
1.8211	1.0	411	1.6256	36.19	19.87	27.87	34.38	24.65	82.13	78.97	80.41	150.43
1.7138	2.0	822	1.5407	36.2	20.04	28.09	34.33	24.4	82.8	78.98	80.74	132.61
1.6039	2.9945	1230	1.5144	36.45	19.67	27.7	34.52	24.99	82.36	79.28	80.68	152.88

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu121
• Datasets 3.3.1
• Tokenizers 0.21.0