Summarizer

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

• Loss: 2.2923
• Rouge1: 0.1987
• Rouge2: 0.0971
• Rougel: 0.1702
• Rougelsum: 0.1701
• Gen Len: 20.0

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	31	2.5664	0.1535	0.0599	0.1259	0.126	20.0
No log	2.0	62	2.5187	0.1742	0.0706	0.1446	0.1446	20.0
No log	3.0	93	2.4849	0.1909	0.0835	0.1607	0.1606	20.0
No log	4.0	124	2.4579	0.197	0.0876	0.1651	0.1651	20.0
No log	5.0	155	2.4365	0.1955	0.086	0.1636	0.1634	20.0
No log	6.0	186	2.4185	0.1969	0.0877	0.1655	0.1654	20.0
No log	7.0	217	2.4042	0.1975	0.0894	0.1669	0.1667	20.0
No log	8.0	248	2.3883	0.1967	0.089	0.1665	0.1664	20.0
No log	9.0	279	2.3775	0.1969	0.0903	0.1672	0.1671	20.0
No log	10.0	310	2.3660	0.1977	0.0913	0.1683	0.1684	20.0
No log	11.0	341	2.3555	0.1976	0.0919	0.1687	0.1687	20.0
No log	12.0	372	2.3491	0.198	0.092	0.1682	0.1682	20.0
No log	13.0	403	2.3410	0.1987	0.0943	0.1692	0.1691	20.0
No log	14.0	434	2.3360	0.1998	0.0957	0.1703	0.1702	20.0
No log	15.0	465	2.3286	0.1998	0.0952	0.1706	0.1706	20.0
No log	16.0	496	2.3226	0.1993	0.095	0.1703	0.1704	20.0
2.4711	17.0	527	2.3194	0.1992	0.0959	0.1707	0.1707	20.0
2.4711	18.0	558	2.3147	0.199	0.0958	0.1708	0.1708	20.0
2.4711	19.0	589	2.3114	0.1987	0.0962	0.1707	0.1708	20.0
2.4711	20.0	620	2.3076	0.199	0.0956	0.1704	0.1703	20.0
2.4711	21.0	651	2.3041	0.1986	0.0963	0.1698	0.1698	20.0
2.4711	22.0	682	2.3012	0.1993	0.0969	0.1707	0.1706	20.0
2.4711	23.0	713	2.2982	0.1993	0.0968	0.1704	0.1704	20.0
2.4711	24.0	744	2.2975	0.1991	0.0965	0.1704	0.1704	20.0
2.4711	25.0	775	2.2968	0.1988	0.0965	0.1701	0.17	20.0
2.4711	26.0	806	2.2951	0.1983	0.0965	0.1701	0.1699	20.0
2.4711	27.0	837	2.2935	0.1986	0.0973	0.1704	0.1702	20.0
2.4711	28.0	868	2.2927	0.1987	0.0971	0.1703	0.1702	20.0
2.4711	29.0	899	2.2925	0.1987	0.0971	0.1702	0.1701	20.0
2.4711	30.0	930	2.2923	0.1987	0.0971	0.1702	0.1701	20.0

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.3.0
• Tokenizers 0.21.0