DanSumT5-baseV_38821

This model is a fine-tuned version of Danish-summarisation/DanSumT5-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.2026
• Rouge1: 34.9358
• Rouge2: 11.6813
• Rougel: 21.4935
• Rougelsum: 27.4979
• Gen Len: 126.3262

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: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• 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	232	2.4684	33.3966	9.9982	19.6472	27.3865	126.8712
No log	2.0	465	2.3905	34.2228	10.5192	20.3584	27.4209	126.8712
2.8064	3.0	697	2.3486	34.5949	11.0682	20.8844	27.3403	126.6738
2.8064	4.0	930	2.3193	34.6865	11.0996	20.9574	27.337	126.2318
2.5767	5.0	1162	2.2963	34.3101	11.0183	20.8461	27.155	126.721
2.5767	6.0	1395	2.2774	34.9299	11.5927	21.3549	27.7805	126.4249
2.483	7.0	1627	2.2646	34.4741	11.1383	21.2722	27.3822	126.3004
2.483	8.0	1860	2.2521	34.9384	11.2651	21.3153	27.5792	126.9828
2.4134	9.0	2092	2.2410	34.9546	11.424	21.1427	27.6608	126.7854
2.4134	10.0	2325	2.2326	34.7566	11.5721	21.4418	27.5167	126.7425
2.3576	11.0	2557	2.2263	34.5968	11.623	21.2384	27.365	126.4506
2.3576	12.0	2790	2.2194	34.7363	11.5612	21.47	27.6572	126.5665
2.3288	13.0	3022	2.2142	34.971	11.7203	21.49	27.7418	126.5665
2.3288	14.0	3255	2.2114	34.761	11.6621	21.3963	27.568	126.6266
2.3288	15.0	3487	2.2064	34.9197	11.5475	21.4017	27.6388	126.3305
2.2951	16.0	3720	2.2067	34.8124	11.615	21.5177	27.605	126.3605
2.2951	17.0	3952	2.2042	34.7608	11.4738	21.3464	27.379	126.4034
2.2832	18.0	4185	2.2032	34.7593	11.6239	21.4029	27.4669	126.2489
2.2832	19.0	4417	2.2029	34.8386	11.5919	21.4719	27.5147	126.2318
2.2571	19.96	4640	2.2026	34.9358	11.6813	21.4935	27.4979	126.3262

Framework versions

• Transformers 4.32.1
• Pytorch 2.1.0
• Datasets 2.12.0
• Tokenizers 0.13.3