LifeScienceBART

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: 4.4310
• Rouge1: 52.3694
• Rouge2: 17.5874
• Rougel: 36.4217
• Rougelsum: 48.765
• Bertscore Precision: 82.295
• Bertscore Recall: 83.951
• Bertscore F1: 83.1121
• Bleu: 0.1308
• Gen Len: 227.8869

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
6.1988	0.0881	100	6.0815	43.5317	12.8172	29.5886	40.6668	78.4798	81.4664	79.9395	0.0939	227.8869
5.7388	0.1762	200	5.6510	41.3899	12.8237	29.1108	38.0304	77.5037	81.7443	79.5601	0.0978	227.8869
5.3718	0.2643	300	5.2822	46.279	14.1045	31.7158	43.1347	79.8268	82.2875	81.0344	0.1041	227.8869
5.1682	0.3524	400	5.1072	48.1957	15.1732	32.7384	44.0672	80.3745	82.94	81.6328	0.1137	227.8869
5.1315	0.4405	500	4.9408	48.9502	15.6058	33.6297	45.5085	81.0706	83.1289	82.0835	0.1158	227.8869
4.9456	0.5286	600	4.7786	48.4843	15.8565	34.014	45.2987	80.9541	83.0806	81.9998	0.1151	227.8869
4.8396	0.6167	700	4.6607	51.3313	16.5503	35.0136	47.9755	82.0251	83.4743	82.7408	0.1210	227.8869
4.7481	0.7048	800	4.5922	51.9257	16.9939	35.583	48.1998	82.2219	83.8107	83.0061	0.1262	227.8869
4.6688	0.7929	900	4.5112	51.3896	17.1313	35.8696	47.7303	81.926	83.7943	82.8465	0.1277	227.8869
4.4321	0.8810	1000	4.4624	52.6168	17.6855	36.2987	49.0759	82.3644	83.8994	83.1222	0.1305	227.8869
4.5732	0.9691	1100	4.4310	52.3694	17.5874	36.4217	48.765	82.295	83.951	83.1121	0.1308	227.8869

Framework versions

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