extract_long_text_unbalanced_smaller_6

This model is a fine-tuned version of weny22/sum_model_t5_saved on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.2687
• Rouge1: 0.2023
• Rouge2: 0.0727
• Rougel: 0.1638
• Rougelsum: 0.1637
• Gen Len: 18.9793

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: 0.002
• train_batch_size: 64
• eval_batch_size: 64
• 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	Gen Len
No log	1.0	335	2.3733	0.1898	0.059	0.1502	0.1501	18.9833
3.2662	2.0	670	2.2397	0.185	0.0609	0.1493	0.1492	18.962
2.392	3.0	1005	2.1809	0.191	0.0655	0.1545	0.1544	18.9607
2.392	4.0	1340	2.1512	0.1923	0.0675	0.1552	0.1551	18.972
2.1606	5.0	1675	2.1182	0.1938	0.0689	0.1571	0.1571	18.974
2.0276	6.0	2010	2.1242	0.1944	0.0671	0.1576	0.1575	18.9733
2.0276	7.0	2345	2.1003	0.1944	0.0683	0.1578	0.1578	18.9833
1.8856	8.0	2680	2.1301	0.1993	0.0718	0.1619	0.162	18.9907
1.7948	9.0	3015	2.1103	0.1978	0.0692	0.1605	0.1605	18.978
1.7948	10.0	3350	2.1220	0.1999	0.0716	0.1628	0.1627	18.9773
1.683	11.0	3685	2.1290	0.1981	0.0699	0.1604	0.1606	18.982
1.6151	12.0	4020	2.1430	0.2024	0.0727	0.1632	0.1632	18.9793
1.6151	13.0	4355	2.1486	0.1976	0.0721	0.1606	0.1607	18.9813
1.5263	14.0	4690	2.1857	0.2032	0.0737	0.1645	0.1645	18.9873
1.4647	15.0	5025	2.2031	0.2029	0.0719	0.1639	0.164	18.9873
1.4647	16.0	5360	2.2044	0.2043	0.0744	0.1659	0.1659	18.9853
1.3972	17.0	5695	2.2325	0.2031	0.0724	0.1637	0.1638	18.9867
1.3577	18.0	6030	2.2473	0.2031	0.0724	0.164	0.1639	18.98
1.3577	19.0	6365	2.2544	0.2029	0.0733	0.1638	0.164	18.984
1.3132	20.0	6700	2.2687	0.2023	0.0727	0.1638	0.1637	18.9793

Framework versions

• Transformers 4.39.1
• Pytorch 2.2.2+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2