mt5-deed-sum

This model is a fine-tuned version of Hasanur525/deed_summarization_mt5_version_1 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4953
• Rouge1: 1.5754
• Rouge2: 1.087
• Rougel: 1.5005
• Rougelsum: 1.4211
• Gen Len: 310.6981

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
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 5000
• num_epochs: 22

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
0.0915	1.0	375	0.5844	0.7311	0.4193	0.7311	0.7311	289.3396
0.9545	2.0	750	0.5858	0.6289	0.444	0.6289	0.6289	291.5912
0.8026	3.0	1125	0.5817	1.1119	0.6733	1.067	1.0428	295.0692
0.2525	4.0	1500	0.5698	0.7311	0.4193	0.7311	0.7311	299.7987
1.5794	5.0	1875	0.5685	0.8096	0.4733	0.7714	0.7549	286.0126
0.0558	6.0	2250	0.5701	0.5003	0.3431	0.5003	0.4785	301.6855
0.4973	7.0	2625	0.5521	1.1281	0.7349	0.9983	0.9983	295.0692
1.1935	8.0	3000	0.5661	1.3444	0.9964	1.2673	1.2213	324.3648
0.0752	9.0	3375	0.5531	1.4883	1.0199	1.4252	1.3979	301.0377
0.216	10.0	3750	0.5573	1.5516	1.0371	1.5047	1.4656	319.195
0.3619	11.0	4125	0.5571	1.2368	0.8055	1.2326	1.2146	294.4717
0.1881	12.0	4500	0.5293	1.2922	0.941	1.2149	1.2084	305.9057
0.2247	13.0	4875	0.5340	1.0581	0.594	0.9989	0.987	306.3774
0.0715	14.0	5250	0.5211	1.2905	0.8861	1.259	1.2143	321.6226
0.1851	15.0	5625	0.5231	1.4625	0.9737	1.3919	1.3637	318.4969
0.5285	16.0	6000	0.5154	1.1892	0.8552	1.1401	1.1061	313.2138
0.0482	17.0	6375	0.5032	1.1826	0.8687	1.1554	1.1554	327.1824
0.0733	18.0	6750	0.5193	1.6133	1.1373	1.5626	1.5085	317.8113
0.2814	19.0	7125	0.5007	1.5689	1.1133	1.5189	1.4606	307.7421
0.0672	20.0	7500	0.4959	1.5754	1.078	1.489	1.4166	316.6164
0.2456	21.0	7875	0.4966	1.5754	1.087	1.5005	1.4211	314.3396
0.0405	22.0	8250	0.4953	1.5754	1.087	1.5005	1.4211	310.6981

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0.dev20230811+cu121
• Datasets 2.17.0
• Tokenizers 0.15.2