mt5-small-finetuned-govreport-summarization

This model is a fine-tuned version of google/mt5-small on the govreport-summarization dataset. It achieves the following results on the evaluation set:

• Loss: 2.9193
• Rouge1: 5.4727
• Rouge2: 1.8064
• Rougel: 4.7904
• Rougelsum: 5.1785

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: 5.6e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 16

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
8.1803	1.0	225	3.4063	4.8262	1.0677	4.1029	4.6438
4.1012	2.0	450	3.2004	4.888	1.2529	4.0737	4.6698
3.8386	3.0	675	3.1341	5.0027	1.1715	4.1397	4.7616
3.6986	4.0	900	3.0698	5.3287	1.6223	4.6697	5.0159
3.6007	5.0	1125	3.0346	5.5318	1.7741	4.8195	5.2351
3.5376	6.0	1350	3.0039	4.5345	1.3055	4.0118	4.3259
3.4794	7.0	1575	2.9845	4.755	1.5096	4.2156	4.5376
3.4373	8.0	1800	2.9699	4.6843	1.409	4.0942	4.4492
3.4007	9.0	2025	2.9569	5.5517	1.8103	4.8226	5.2639
3.3788	10.0	2250	2.9415	5.4873	1.8689	4.8027	5.2162
3.3549	11.0	2475	2.9429	5.3814	1.7672	4.7337	5.1079
3.3386	12.0	2700	2.9338	5.4238	1.7718	4.7339	5.1216
3.3195	13.0	2925	2.9224	5.4666	1.8941	4.79	5.1824
3.311	14.0	3150	2.9223	5.4197	1.7975	4.7752	5.1176
3.3027	15.0	3375	2.9202	5.494	1.8446	4.7876	5.1981
3.2961	16.0	3600	2.9193	5.4727	1.8064	4.7904	5.1785

Framework versions

• Transformers 4.42.3
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1