mt5-large-finetuned-scope-summarization

This model is a fine-tuned version of google/mt5-large on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 8.2775
• Rouge1: 5.918
• Rouge2: 1.0667
• Rougel: 5.7247
• Rougelsum: 5.552

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: 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
27.8719	1.0	13	15.8303	9.9779	0.8912	8.8304	8.8653
25.4142	2.0	26	20.3410	11.3301	1.0662	9.8807	9.8442
24.8026	3.0	39	16.5876	11.1912	1.5008	9.9776	9.9685
23.7918	4.0	52	14.0667	11.5953	1.6391	10.2961	10.1512
21.945	5.0	65	12.3075	10.6522	1.2121	10.0748	10.0261
18.8588	6.0	78	11.8270	11.4944	1.4152	9.9891	9.9505
16.587	7.0	91	10.7425	9.9989	1.425	8.9661	8.9811
15.9949	8.0	104	10.2228	10.0086	1.6533	8.9911	9.0047
15.2301	9.0	117	11.2979	9.2011	1.425	8.9267	8.8763
14.9655	10.0	130	11.3654	9.3934	1.6533	8.9243	8.8443
14.7982	11.0	143	10.7718	8.5085	1.4133	8.0936	8.0127
13.5222	12.0	156	10.0961	7.849	1.1637	7.3283	7.1943
13.0959	13.0	169	9.4677	8.0846	1.1637	7.1215	7.0501
13.0554	14.0	182	8.9576	7.0454	1.2494	6.7761	6.6897
13.1098	15.0	195	8.7926	7.9192	1.4133	7.742	7.6718
12.4133	16.0	208	8.5472	7.0176	1.2819	6.8465	6.8276
12.4751	17.0	221	8.5494	5.918	1.0667	5.7247	5.552
11.9681	18.0	234	8.5223	5.918	1.0667	5.7247	5.552
11.8797	19.0	247	8.3327	5.918	1.0667	5.7247	5.552
11.8815	20.0	260	8.2775	5.918	1.0667	5.7247	5.552

Framework versions

• Transformers 4.38.1
• Pytorch 2.2.1+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2