sum_model_lr2e_3_20epoch

This model is a fine-tuned version of weny22/sum_model_t5_saved on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.0536
• Rouge1: 0.2178
• Rouge2: 0.0887
• Rougel: 0.1786
• Rougelsum: 0.1788
• Gen Len: 18.982

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: 0.002
• train_batch_size: 64
• eval_batch_size: 64
• 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	335	2.1360	0.1968	0.0675	0.1569	0.1571	18.9787
2.7013	2.0	670	2.0015	0.2028	0.0743	0.1642	0.1644	18.9833
2.149	3.0	1005	1.9683	0.2131	0.083	0.1727	0.173	18.974
2.149	4.0	1340	1.9278	0.2094	0.0829	0.1705	0.1708	18.9753
1.929	5.0	1675	1.9005	0.2159	0.0861	0.1766	0.1767	18.9753
1.8038	6.0	2010	1.9047	0.2155	0.086	0.1745	0.1747	18.9867
1.8038	7.0	2345	1.9229	0.2147	0.0869	0.1746	0.1749	18.9887
1.6671	8.0	2680	1.8959	0.214	0.0866	0.1757	0.1761	18.9747
1.5832	9.0	3015	1.9048	0.2163	0.0865	0.176	0.1762	18.98
1.5832	10.0	3350	1.8947	0.217	0.0871	0.1769	0.1771	18.984
1.4739	11.0	3685	1.9152	0.2147	0.0882	0.1766	0.1769	18.97
1.4101	12.0	4020	1.9340	0.2148	0.0876	0.1767	0.1769	18.986
1.4101	13.0	4355	1.9522	0.2136	0.0857	0.1742	0.1745	18.9887
1.3249	14.0	4690	1.9670	0.2187	0.0902	0.1792	0.1795	18.9893
1.2673	15.0	5025	1.9860	0.2169	0.0881	0.1782	0.1785	18.9947
1.2673	16.0	5360	1.9971	0.2162	0.0867	0.1777	0.178	18.9673
1.2021	17.0	5695	2.0146	0.2163	0.0877	0.177	0.1773	18.9893
1.1621	18.0	6030	2.0313	0.2171	0.0887	0.1778	0.1779	18.986
1.1621	19.0	6365	2.0466	0.2182	0.0893	0.1791	0.1792	18.9873
1.1192	20.0	6700	2.0536	0.2178	0.0887	0.1786	0.1788	18.982

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2