bart-base-summarization-medical_on_cnn-49

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: 3.3782
• Rouge1: 0.2538
• Rouge2: 0.0951
• Rougel: 0.1997
• Rougelsum: 0.2242
• Gen Len: 18.562

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: 49
• 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
2.699	1.0	1250	3.3753	0.2516	0.0907	0.1966	0.2214	19.026
2.6011	2.0	2500	3.3638	0.2505	0.0913	0.1968	0.2211	18.839
2.578	3.0	3750	3.3738	0.2516	0.0918	0.1971	0.2208	18.888
2.532	4.0	5000	3.3729	0.2523	0.0946	0.1993	0.223	18.506
2.5583	5.0	6250	3.3794	0.2528	0.0939	0.1984	0.2233	18.612
2.539	6.0	7500	3.3782	0.2538	0.0951	0.1997	0.2242	18.562

Framework versions

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