2_smtg

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

• Loss: 1.9346
• Rouge1: 0.1982
• Rouge2: 0.1052
• Rougel: 0.1709
• Rougelsum: 0.1711
• Gen Len: 19.0

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	124	2.2154	0.1881	0.0892	0.1571	0.157	18.996
No log	2.0	248	2.1455	0.2003	0.1039	0.1695	0.1696	19.0
No log	3.0	372	2.0963	0.2011	0.1043	0.1706	0.1706	19.0
No log	4.0	496	2.0696	0.2014	0.105	0.1708	0.1708	19.0
2.4198	5.0	620	2.0437	0.1991	0.1016	0.1693	0.1694	19.0
2.4198	6.0	744	2.0256	0.1983	0.1016	0.1694	0.1695	19.0
2.4198	7.0	868	2.0109	0.2003	0.1044	0.1702	0.1705	19.0
2.4198	8.0	992	1.9969	0.1981	0.1025	0.1692	0.1694	19.0
2.2056	9.0	1116	1.9849	0.1984	0.103	0.1696	0.1699	19.0
2.2056	10.0	1240	1.9738	0.1985	0.1032	0.1702	0.1704	19.0
2.2056	11.0	1364	1.9661	0.1976	0.1029	0.1694	0.1697	19.0
2.2056	12.0	1488	1.9591	0.1986	0.1038	0.1704	0.1706	19.0
2.1209	13.0	1612	1.9535	0.1994	0.1045	0.1708	0.1709	19.0
2.1209	14.0	1736	1.9486	0.1986	0.1047	0.1706	0.1708	19.0
2.1209	15.0	1860	1.9440	0.1988	0.1053	0.1709	0.1711	19.0
2.1209	16.0	1984	1.9406	0.1983	0.1057	0.1708	0.1709	19.0
2.0754	17.0	2108	1.9378	0.199	0.1062	0.1712	0.1712	19.0
2.0754	18.0	2232	1.9361	0.1986	0.1057	0.1713	0.1714	19.0
2.0754	19.0	2356	1.9348	0.1986	0.1056	0.1712	0.1713	19.0
2.0754	20.0	2480	1.9346	0.1982	0.1052	0.1709	0.1711	19.0

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1.post200
• Datasets 2.10.0
• Tokenizers 0.13.2