fine-tuned-BioBART-2048-inputs-10-epochs

This model is a fine-tuned version of GanjinZero/biobart-v2-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.7099
• Rouge1: 0.2904
• Rouge2: 0.1173
• Rougel: 0.2687
• Rougelsum: 0.2692
• Gen Len: 14.66

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: 2e-05
• 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
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	151	0.7536	0.2059	0.0784	0.1881	0.1881	13.31
No log	2.0	302	0.7161	0.2569	0.0831	0.2279	0.2278	13.88
No log	3.0	453	0.7013	0.2322	0.0818	0.2055	0.2059	14.57
0.7283	4.0	604	0.6976	0.2835	0.1095	0.2585	0.2584	14.34
0.7283	5.0	755	0.7012	0.2749	0.0921	0.2521	0.2528	14.35
0.7283	6.0	906	0.6963	0.2957	0.1073	0.2688	0.269	14.97
0.5246	7.0	1057	0.7043	0.2824	0.1067	0.257	0.257	14.68
0.5246	8.0	1208	0.7043	0.292	0.1158	0.2706	0.2722	14.16
0.5246	9.0	1359	0.7080	0.2849	0.1087	0.2603	0.2615	14.69
0.4414	10.0	1510	0.7099	0.2904	0.1173	0.2687	0.2692	14.66

Framework versions

• Transformers 4.36.2
• Pytorch 1.12.1+cu113
• Datasets 2.15.0
• Tokenizers 0.15.0