t5-small-finetuned-xsum

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

• Loss: 2.9389
• Rouge1: 0.2199
• Rouge2: 0.0413
• Rougel: 0.1739
• Rougelsum: 0.1836

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: 0.0002
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
3.9818	1.0	1	3.5789	0.1857	0.0245	0.1420	0.1556
3.5098	2.0	2	3.4107	0.1863	0.0245	0.1391	0.1564
3.1669	3.0	3	3.2830	0.2008	0.0254	0.1466	0.1703
2.8568	4.0	4	3.1945	0.1980	0.0222	0.1411	0.1622
2.7102	5.0	5	3.1215	0.2019	0.0222	0.1472	0.1609
2.4563	6.0	6	3.0798	0.2167	0.0189	0.1533	0.1737
2.3367	7.0	7	3.0364	0.2050	0.0139	0.1420	0.1577
2.269	8.0	8	3.0071	0.2041	0.0139	0.1435	0.1561
2.0398	9.0	9	2.9865	0.2246	0.0139	0.1510	0.1721
1.9314	10.0	10	2.9783	0.2076	0.0139	0.1542	0.1681
1.9148	11.0	11	2.9684	0.2076	0.0139	0.1542	0.1681
1.8131	12.0	12	2.9598	0.2076	0.0139	0.1542	0.1681
1.7866	13.0	13	2.9497	0.2195	0.0184	0.1501	0.1722
1.689	14.0	14	2.9451	0.2067	0.0203	0.1453	0.1621
1.7257	15.0	15	2.9405	0.2155	0.0321	0.1599	0.1777
1.6441	16.0	16	2.9405	0.2155	0.0321	0.1599	0.1777
1.574	17.0	17	2.9409	0.2155	0.0321	0.1599	0.1777
1.587	18.0	18	2.9393	0.2260	0.0388	0.1678	0.1860
1.5362	19.0	19	2.9387	0.2199	0.0413	0.1739	0.1836
1.5133	20.0	20	2.9389	0.2199	0.0413	0.1739	0.1836

Framework versions

• Transformers 4.39.3
• Pytorch 2.2.2+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2