deed_summarization_mt5_version_1

This model is a fine-tuned version of google/mt5-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5863
• Rouge1: 1.0138
• Rouge2: 0.6875
• Rougel: 1.0233
• Rougelsum: 1.0941
• Gen Len: 288.1509

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: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
25.5422	1.0	375	16.6467	0.6477	0.0	0.6383	0.6639	25.0
11.9626	2.0	750	13.5214	0.6633	0.0	0.6553	0.6745	38.5912
10.1002	3.0	1125	7.4294	0.7257	0.0	0.7163	0.7264	386.6164
2.8844	4.0	1500	2.8574	0.7257	0.0	0.7163	0.7264	499.0
4.1183	5.0	1875	9.6893	0.7257	0.0	0.7163	0.7264	499.0
1.4443	6.0	2250	2.4224	0.7257	0.0	0.7163	0.7264	466.7673
3.8512	7.0	2625	1.5813	0.7257	0.0	0.7163	0.7264	432.4717
8.6527	8.0	3000	1.4532	0.7257	0.0	0.7163	0.7264	480.6164
0.5302	9.0	3375	1.1597	0.7257	0.0	0.7163	0.7264	419.239
1.2311	10.0	3750	0.9806	0.9895	0.1006	0.9135	0.9189	383.6855
0.8903	11.0	4125	0.8961	0.9609	0.1578	0.8871	0.8934	376.6038
0.8742	12.0	4500	0.8109	1.1104	0.2243	1.0038	1.007	388.3648
0.5934	13.0	4875	0.7588	0.3145	0.2287	0.3145	0.3145	341.717
0.1715	14.0	5250	0.7073	0.2795	0.2013	0.2795	0.2795	333.434
0.4363	15.0	5625	0.6780	0.4368	0.2287	0.4368	0.4368	326.7044
1.0736	16.0	6000	0.6647	0.7163	0.5169	0.7512	0.7512	299.4151
0.1069	17.0	6375	0.6294	0.856	0.6038	0.863	0.8595	312.434
0.1434	18.0	6750	0.6358	0.7512	0.5222	0.7862	0.808	291.4403
0.4344	19.0	7125	0.6164	1.1082	0.7576	1.1305	1.1574	304.7484
0.1038	20.0	7500	0.6066	0.8572	0.6108	0.8758	0.9085	297.3145
0.5519	21.0	7875	0.5972	0.4354	0.2935	0.5382	0.5382	281.5786
0.0804	22.0	8250	0.5994	0.6464	0.5583	0.7741	0.7794	305.805
0.3696	23.0	8625	0.5884	0.6362	0.3246	0.6362	0.6362	291.434
0.3966	24.0	9000	0.5852	0.7133	0.408	0.7311	0.8082	281.9119
0.3484	25.0	9375	0.5863	1.0138	0.6875	1.0233	1.0941	288.1509

Framework versions

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