fine-tuned-BioBART-15-epochs-1024-input-128-output

This model is a fine-tuned version of GanjinZero/biobart-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.5422
• Rouge1: 0.1912
• Rouge2: 0.042
• Rougel: 0.1492
• Rougelsum: 0.15
• Gen Len: 29.97

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	151	5.4038	0.0026	0.0006	0.0024	0.0024	5.98
No log	2.0	302	1.8682	0.0475	0.0066	0.0452	0.043	6.59
No log	3.0	453	1.6512	0.0753	0.0169	0.0587	0.0586	22.62
4.1375	4.0	604	1.5702	0.1472	0.0366	0.1124	0.1113	42.92
4.1375	5.0	755	1.5256	0.167	0.0337	0.1309	0.1305	45.89
4.1375	6.0	906	1.5057	0.1435	0.0305	0.1132	0.1134	32.45
1.1893	7.0	1057	1.4854	0.1655	0.0388	0.129	0.1295	34.34
1.1893	8.0	1208	1.4845	0.1635	0.0423	0.1238	0.1252	37.77
1.1893	9.0	1359	1.4980	0.1712	0.0363	0.1382	0.1388	29.68
0.8262	10.0	1510	1.5052	0.1917	0.0431	0.1486	0.1497	32.88
0.8262	11.0	1661	1.5167	0.1731	0.0374	0.1402	0.1403	29.9
0.8262	12.0	1812	1.5267	0.1675	0.035	0.1335	0.1337	29.35
0.8262	13.0	1963	1.5329	0.1839	0.0401	0.1465	0.1465	28.23
0.61	14.0	2114	1.5440	0.1904	0.0452	0.1522	0.1527	29.33
0.61	15.0	2265	1.5422	0.1912	0.042	0.1492	0.15	29.97

Framework versions

• Transformers 4.36.2
• Pytorch 1.12.1+cu113
• Datasets 2.16.1
• Tokenizers 0.15.0