t5-small-finetuned-xsum

This model is a fine-tuned version of t5-small on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0734
• Rouge1: 99.9038
• Rouge2: 99.838
• Rougel: 99.9145
• Rougelsum: 99.9038
• Gen Len: 93.9181

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 40
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	180	1.7815	9.7268	2.7047	8.7069	8.7035	155.8472
No log	2.0	360	0.6270	28.7135	19.99	27.1646	27.1386	265.2903
2.122	3.0	540	0.3572	21.4211	17.5143	21.0387	20.9118	142.7333
2.122	4.0	720	0.2757	92.8223	90.5077	92.0061	92.0015	87.0847
2.122	5.0	900	0.2493	95.6972	94.5082	95.5057	95.522	91.8556
0.4002	6.0	1080	0.2348	96.8942	96.2704	96.7552	96.7736	96.0764
0.4002	7.0	1260	0.2227	97.7669	97.4255	97.6867	97.6913	93.9097
0.4002	8.0	1440	0.2111	98.7823	98.5538	98.7622	98.7722	94.2875
0.2717	9.0	1620	0.1979	99.7455	99.6524	99.7428	99.7449	93.8569
0.2717	10.0	1800	0.1843	99.8967	99.8175	99.8953	99.8939	93.875
0.2717	11.0	1980	0.1716	99.9078	99.8578	99.9114	99.9095	93.8556
0.2244	12.0	2160	0.1606	99.9371	99.8807	99.9373	99.9373	93.9236
0.2244	13.0	2340	0.1512	99.9112	99.8535	99.9141	99.9103	93.8542
0.19	14.0	2520	0.1424	99.9573	99.919	99.9573	99.9573	93.9236
0.19	15.0	2700	0.1353	99.9679	99.9421	99.9679	99.9679	93.925
0.19	16.0	2880	0.1290	99.9234	99.8727	99.9323	99.9234	93.8736
0.1652	17.0	3060	0.1235	99.9252	99.8727	99.9359	99.9252	93.9222
0.1652	18.0	3240	0.1184	99.9038	99.8373	99.911	99.9021	93.8722
0.1652	19.0	3420	0.1137	99.9466	99.9074	99.9573	99.9466	93.9236
0.1471	20.0	3600	0.1092	99.9252	99.8727	99.9359	99.9252	93.9222
0.1471	21.0	3780	0.1053	99.9252	99.8727	99.9359	99.9252	93.9222
0.1471	22.0	3960	0.1014	99.9252	99.8727	99.9359	99.9252	93.9222
0.1331	23.0	4140	0.0982	99.9252	99.8727	99.9359	99.9252	93.9222
0.1331	24.0	4320	0.0949	99.9252	99.8727	99.9359	99.9252	93.9208
0.1226	25.0	4500	0.0918	99.9252	99.8727	99.9359	99.9252	93.9208
0.1226	26.0	4680	0.0892	99.9252	99.8727	99.9359	99.9252	93.9208
0.1226	27.0	4860	0.0867	99.9252	99.8727	99.9359	99.9252	93.9208
0.114	28.0	5040	0.0848	99.9145	99.8495	99.9252	99.9145	93.9194
0.114	29.0	5220	0.0828	99.9038	99.838	99.9145	99.9038	93.9181
0.114	30.0	5400	0.0811	99.9145	99.8495	99.9252	99.9145	93.9194
0.1074	31.0	5580	0.0794	99.9038	99.838	99.9145	99.9038	93.9181
0.1074	32.0	5760	0.0781	99.9038	99.838	99.9145	99.9038	93.9181
0.1074	33.0	5940	0.0769	99.9252	99.8669	99.9252	99.9252	93.9194
0.1027	34.0	6120	0.0757	99.9038	99.838	99.9145	99.9038	93.9181
0.1027	35.0	6300	0.0751	99.9038	99.838	99.9145	99.9038	93.9181
0.1027	36.0	6480	0.0745	99.9038	99.838	99.9145	99.9038	93.9181
0.0994	37.0	6660	0.0740	99.9038	99.838	99.9145	99.9038	93.9181
0.0994	38.0	6840	0.0737	99.9038	99.838	99.9145	99.9038	93.9181
0.0975	39.0	7020	0.0735	99.9038	99.838	99.9145	99.9038	93.9181
0.0975	40.0	7200	0.0734	99.9038	99.838	99.9145	99.9038	93.9181

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.0
• Tokenizers 0.15.2