bart-large-pubmedd

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: 2.5626
• Rouge1: 37.72
• Rouge2: 14.01
• Rougel: 22.84
• Rougelsum: 33.26

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: 5e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 1000

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
3.1141	0.2668	200	2.7332	36.95	13.98	22.78	32.78
2.9621	0.5337	400	2.6831	38.12	14.37	22.8	33.44
2.9085	0.8005	600	2.6033	37.28	13.84	22.44	32.9
2.7868	1.0674	800	2.5722	37.18	13.95	22.58	32.94
2.8079	1.3342	1000	2.5626	37.72	14.01	22.84	33.26

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.2.0
• Tokenizers 0.19.1