t5-small-finetuned-dialogsum-v3

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

• Loss: 1.2045
• Rouge1: 38.3615
• Rouge2: 16.0241
• Rougel: 32.901
• Rougelsum: 34.8687
• Gen Len: 18.892

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: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
1.7344	1.0	779	1.4251	33.4125	10.7502	28.0588	30.0903	18.858
1.4975	2.0	1558	1.3623	34.4069	11.9728	29.0576	31.156	18.874
1.4621	3.0	2337	1.3355	34.9786	12.314	29.4869	31.4407	18.86
1.4149	4.0	3116	1.3119	35.5881	12.9123	30.1883	32.0652	18.874
1.4009	5.0	3895	1.2905	36.3104	13.8382	30.893	32.7095	18.882
1.3709	6.0	4674	1.2736	36.3456	13.8426	30.7526	32.6784	18.906
1.3589	7.0	5453	1.2671	36.6543	14.2334	30.98	32.9241	18.892
1.3373	8.0	6232	1.2557	37.2264	14.7072	31.413	33.2844	18.914
1.3168	9.0	7011	1.2520	37.315	14.8744	31.6711	33.4863	18.862
1.3044	10.0	7790	1.2454	37.8787	15.4762	32.3244	34.107	18.886
1.2915	11.0	8569	1.2380	38.0242	15.5379	32.4465	34.292	18.862
1.2926	12.0	9348	1.2362	37.82	15.4074	32.0479	33.9622	18.882
1.2818	13.0	10127	1.2318	38.2168	16.0879	32.592	34.5757	18.892
1.2766	14.0	10906	1.2257	38.559	16.2997	32.9956	34.9149	18.864
1.2666	15.0	11685	1.2245	38.1764	15.9612	32.525	34.6476	18.878
1.2602	16.0	12464	1.2191	38.3852	16.085	32.809	34.7302	18.884
1.2523	17.0	13243	1.2164	38.426	16.1149	32.6806	34.7701	18.894
1.2466	18.0	14022	1.2142	38.6658	16.0599	32.9194	34.905	18.89
1.2332	19.0	14801	1.2152	38.4253	15.9033	32.7993	34.8635	18.896
1.2344	20.0	15580	1.2093	38.6261	16.0519	33.1192	34.9215	18.918
1.2278	21.0	16359	1.2091	38.6618	16.2012	33.134	35.0842	18.904
1.2255	22.0	17138	1.2077	38.6482	16.142	33.0472	35.037	18.906
1.2305	23.0	17917	1.2068	38.6584	16.1184	32.9757	34.9885	18.89
1.2275	24.0	18696	1.2069	38.3795	16.0471	32.9456	34.8267	18.874
1.2227	25.0	19475	1.2064	38.4788	16.1603	33.0022	34.8844	18.87
1.218	26.0	20254	1.2051	38.5133	16.0813	33.0334	34.9492	18.89
1.2183	27.0	21033	1.2046	38.3323	15.839	32.7421	34.7147	18.884
1.2195	28.0	21812	1.2040	38.3573	16.0328	32.86	34.8107	18.892
1.2145	29.0	22591	1.2045	38.3932	16.1115	32.9154	34.8664	18.894
1.212	30.0	23370	1.2045	38.3615	16.0241	32.901	34.8687	18.892

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1