fine-tuned-BioBART-15-epochs-1500-input-256-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: 0.8909
• Rouge1: 0.1801
• Rouge2: 0.0468
• Rougel: 0.1405
• Rougelsum: 0.1403
• Gen Len: 36.53

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: 4
• eval_batch_size: 4
• 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	301	1.4869	0.086	0.0035	0.0743	0.0744	10.36
4.753	2.0	602	0.9880	0.0958	0.0233	0.0687	0.0686	47.88
4.753	3.0	903	0.9108	0.0622	0.009	0.0617	0.0622	10.39
0.9239	4.0	1204	0.8681	0.1159	0.0272	0.085	0.0849	27.95
0.7513	5.0	1505	0.8490	0.1732	0.0333	0.1345	0.1352	37.9
0.7513	6.0	1806	0.8486	0.1523	0.0297	0.12	0.1209	33.96
0.5751	7.0	2107	0.8402	0.1825	0.0426	0.1433	0.1437	40.27
0.5751	8.0	2408	0.8500	0.1712	0.0303	0.1425	0.1427	31.42
0.4702	9.0	2709	0.8542	0.1764	0.0325	0.1275	0.1273	44.93
0.3782	10.0	3010	0.8615	0.1667	0.042	0.1328	0.1335	36.86
0.3782	11.0	3311	0.8714	0.1756	0.0358	0.1364	0.1359	35.21
0.3005	12.0	3612	0.8772	0.1801	0.0368	0.1427	0.1427	33.99
0.3005	13.0	3913	0.8818	0.1685	0.0397	0.1323	0.1331	34.67
0.2417	14.0	4214	0.8891	0.189	0.0495	0.145	0.1445	36.0
0.2148	15.0	4515	0.8909	0.1801	0.0468	0.1405	0.1403	36.53

Framework versions

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