HealthScienceBART

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

• Loss: 3.7248
• Rouge1: 59.8432
• Rouge2: 25.926
• Rougel: 44.3683
• Rougelsum: 56.3382
• Bertscore Precision: 84.199
• Bertscore Recall: 85.5429
• Bertscore F1: 84.8633
• Bleu: 0.2087
• Gen Len: 234.8216

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: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Bertscore Precision	Bertscore Recall	Bertscore F1	Bleu	Gen Len
5.662	0.0826	100	5.4864	49.8946	18.6145	35.6824	47.1811	80.6966	82.5402	81.6048	0.1476	234.8216
5.2036	0.1653	200	4.9823	52.1848	20.4176	37.3029	48.9924	81.1422	83.2665	82.1871	0.1634	234.8216
4.7061	0.2479	300	4.6422	54.5492	21.4905	38.8501	51.1097	82.0428	83.8584	82.9376	0.1730	234.8216
4.657	0.3305	400	4.4252	54.072	22.1609	39.6324	50.5966	81.9494	84.1622	83.0371	0.1793	234.8216
4.3613	0.4131	500	4.2631	56.8149	23.0471	40.9892	53.0419	83.0301	84.669	83.8388	0.1871	234.8216
4.2804	0.4958	600	4.1142	56.8254	23.7321	41.7326	52.8585	82.8372	84.8241	83.8154	0.1915	234.8216
4.2477	0.5784	700	3.9926	57.2046	23.9303	42.3439	53.6018	83.216	84.9845	84.0878	0.1929	234.8216
4.1188	0.6610	800	3.9193	57.9987	24.8441	43.1811	54.4399	83.6075	85.2031	84.395	0.1999	234.8216
3.8678	0.7436	900	3.8320	59.1683	25.1465	43.4643	55.6762	83.9212	85.315	84.6099	0.2019	234.8216
3.8831	0.8263	1000	3.7889	59.3948	25.4051	43.821	55.8124	84.0802	85.4569	84.7606	0.2044	234.8216
3.7856	0.9089	1100	3.7498	59.535	25.6124	44.1831	56.071	84.0653	85.4796	84.7641	0.2063	234.8216
3.8875	0.9915	1200	3.7248	59.8432	25.926	44.3683	56.3382	84.199	85.5429	84.8633	0.2087	234.8216

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1