extract_long_text_unbalanced_smaller_original_text_4

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: 2.4822
• Rouge1: 0.1996
• Rouge2: 0.0696
• Rougel: 0.1603
• Rougelsum: 0.1603
• Gen Len: 18.9893

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	72	2.6068	0.1714	0.0484	0.1369	0.1369	18.988
No log	2.0	144	2.3827	0.1803	0.0547	0.1414	0.1413	18.994
No log	3.0	216	2.2953	0.1858	0.0568	0.1457	0.1458	19.0
No log	4.0	288	2.2509	0.1879	0.0598	0.1479	0.1478	18.9953
No log	5.0	360	2.2338	0.1837	0.0568	0.1448	0.1449	18.9967
No log	6.0	432	2.2428	0.1869	0.0608	0.1484	0.1484	18.9953
3.0458	7.0	504	2.2195	0.1927	0.0628	0.1537	0.1536	18.9867
3.0458	8.0	576	2.2549	0.1933	0.0619	0.152	0.1522	18.9967
3.0458	9.0	648	2.2675	0.1953	0.0643	0.156	0.156	18.9607
3.0458	10.0	720	2.2858	0.198	0.0665	0.1572	0.1573	18.9807
3.0458	11.0	792	2.2980	0.1943	0.0653	0.1555	0.1555	18.972
3.0458	12.0	864	2.3413	0.1998	0.0683	0.1596	0.1595	18.9807
3.0458	13.0	936	2.3324	0.1988	0.0677	0.1587	0.1585	18.9733
1.907	14.0	1008	2.3481	0.2002	0.0688	0.1598	0.1599	18.9913
1.907	15.0	1080	2.4027	0.2024	0.0705	0.1616	0.1616	18.9887
1.907	16.0	1152	2.4132	0.2031	0.0728	0.1633	0.1634	18.9833
1.907	17.0	1224	2.4393	0.1988	0.0683	0.1584	0.1584	18.9853
1.907	18.0	1296	2.4435	0.199	0.0699	0.1591	0.1592	18.9867
1.907	19.0	1368	2.4703	0.2013	0.0704	0.1606	0.1608	18.9873
1.907	20.0	1440	2.4822	0.1996	0.0696	0.1603	0.1603	18.9893

Framework versions

• Transformers 4.38.2
• Pytorch 2.1.2+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2