tanatapanun/fine-tuned-BioBART-2048-inputs-20-epochs

fine-tuned-BART-2048-inputs-20-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: 0.7640
• Rouge1: 0.318
• Rouge2: 0.1243
• Rougel: 0.2884
• Rougelsum: 0.2894
• Gen Len: 15.42

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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	151	0.7532	0.2007	0.0751	0.1827	0.1821	13.29
No log	2.0	302	0.7148	0.261	0.0836	0.2299	0.2312	13.92
No log	3.0	453	0.6995	0.248	0.0862	0.2195	0.2201	14.49
0.724	4.0	604	0.6956	0.2944	0.1061	0.2658	0.2665	14.31
0.724	5.0	755	0.7029	0.3061	0.1203	0.2808	0.283	14.81
0.724	6.0	906	0.6965	0.2848	0.1118	0.2596	0.2584	15.0
0.5016	7.0	1057	0.7097	0.2874	0.1207	0.2558	0.2562	15.0
0.5016	8.0	1208	0.7140	0.293	0.1143	0.2617	0.2641	14.3
0.5016	9.0	1359	0.7191	0.3198	0.1222	0.2877	0.2903	14.75
0.3838	10.0	1510	0.7274	0.3127	0.1265	0.2863	0.2874	14.82
0.3838	11.0	1661	0.7312	0.3129	0.1282	0.2821	0.2819	14.97
0.3838	12.0	1812	0.7419	0.2974	0.1123	0.2726	0.2725	14.98
0.3838	13.0	1963	0.7441	0.2945	0.1139	0.2682	0.2681	15.1
0.3153	14.0	2114	0.7490	0.2969	0.1207	0.2743	0.2753	15.29
0.3153	15.0	2265	0.7536	0.2971	0.1116	0.2674	0.2689	14.83
0.3153	16.0	2416	0.7564	0.301	0.1078	0.271	0.2726	15.3
0.2646	17.0	2567	0.7585	0.2989	0.1117	0.2737	0.2744	15.21
0.2646	18.0	2718	0.7630	0.2944	0.1078	0.2641	0.265	15.12
0.2646	19.0	2869	0.7632	0.2986	0.1089	0.2669	0.2683	15.25
0.2428	20.0	3020	0.7640	0.318	0.1243	0.2884	0.2894	15.42

Framework versions

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