my_summarization_model

This model is a fine-tuned version of google-t5/t5-small on an billsum dataset. It achieves the following results on the evaluation set:

• Loss: 2.2280
• Rouge1: 0.4067
• Rouge2: 0.1832
• Rougel: 0.2719
• Rougelsum: 0.2717
• Gen Len: 126.8427

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: 25
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	124	2.5595	0.3549	0.1384	0.2234	0.2233	123.0242
No log	2.0	248	2.4759	0.3779	0.1517	0.2462	0.2461	124.0323
No log	3.0	372	2.4305	0.3921	0.1647	0.2583	0.2582	126.379
No log	4.0	496	2.3922	0.393	0.1666	0.2609	0.261	126.1089
2.651	5.0	620	2.3726	0.3956	0.1689	0.2637	0.2641	126.3831
2.651	6.0	744	2.3473	0.3985	0.1736	0.2666	0.2669	126.4153
2.651	7.0	868	2.3269	0.3991	0.1717	0.2651	0.2651	126.4315
2.651	8.0	992	2.3154	0.3964	0.1695	0.2648	0.2647	126.5161
2.4496	9.0	1116	2.3047	0.4022	0.1755	0.2695	0.2694	126.5726
2.4496	10.0	1240	2.2988	0.4021	0.1758	0.27	0.2699	126.5161
2.4496	11.0	1364	2.2797	0.4033	0.1779	0.2718	0.2716	126.5726
2.4496	12.0	1488	2.2765	0.4072	0.1804	0.2719	0.2718	126.4758
2.3631	13.0	1612	2.2661	0.4074	0.1797	0.2722	0.2723	126.6452
2.3631	14.0	1736	2.2585	0.4042	0.1769	0.27	0.2698	126.6089
2.3631	15.0	1860	2.2539	0.4066	0.1797	0.2721	0.2722	126.6613
2.3631	16.0	1984	2.2497	0.403	0.176	0.2696	0.2697	126.6371
2.3203	17.0	2108	2.2438	0.4038	0.1783	0.2706	0.2707	126.7339
2.3203	18.0	2232	2.2375	0.4034	0.1787	0.2691	0.2693	126.7903
2.3203	19.0	2356	2.2354	0.4016	0.1779	0.2676	0.2677	126.8427
2.3203	20.0	2480	2.2334	0.4041	0.1787	0.2697	0.2697	126.8952
2.285	21.0	2604	2.2315	0.4026	0.1797	0.2694	0.2693	126.7903
2.285	22.0	2728	2.2302	0.4044	0.1804	0.27	0.27	126.7903
2.285	23.0	2852	2.2284	0.4055	0.1827	0.2716	0.2714	126.7379
2.285	24.0	2976	2.2283	0.4061	0.1825	0.2716	0.2715	126.7903
2.2698	25.0	3100	2.2280	0.4067	0.1832	0.2719	0.2717	126.8427

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.19.1