pegasus-xsum-clara-med

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

• Loss: 1.9013
• Rouge1: 43.7595
• Rouge2: 25.7022
• Rougel: 39.6153
• Rougelsum: 39.7151

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: 5.6e-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

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
No log	1.0	190	2.5468	41.6125	24.1264	37.7704	37.8615
No log	2.0	380	2.3603	41.9598	24.315	38.1087	38.217
2.7787	3.0	570	2.2604	42.0463	24.5067	38.1632	38.2716
2.7787	4.0	760	2.1846	42.1471	24.639	38.3677	38.471
2.2691	5.0	950	2.1361	42.4562	24.8962	38.6107	38.7065
2.2691	6.0	1140	2.0887	42.6005	24.947	38.7049	38.805
2.2691	7.0	1330	2.0617	42.7946	24.9509	38.9123	39.0003
2.0313	8.0	1520	2.0222	43.0201	25.3552	39.151	39.266
2.0313	9.0	1710	2.0049	43.2293	25.4719	39.4239	39.4944
1.872	10.0	1900	1.9899	43.2629	25.5285	39.4124	39.4591
1.872	11.0	2090	1.9772	43.4294	25.8006	39.5863	39.6726
1.872	12.0	2280	1.9630	43.63	25.7259	39.5521	39.6888
1.7497	13.0	2470	1.9513	43.4053	25.5567	39.4567	39.5918
1.7497	14.0	2660	1.9336	43.2584	25.4554	39.2917	39.3944
1.6609	15.0	2850	1.9345	43.2644	25.5958	39.3474	39.4645
1.6609	16.0	3040	1.9152	43.4404	25.6127	39.4472	39.5418
1.6609	17.0	3230	1.9106	43.2751	25.3213	39.2723	39.3871
1.5809	18.0	3420	1.9125	43.2335	25.341	39.2705	39.3577
1.5809	19.0	3610	1.9086	43.1679	25.3275	39.1858	39.303
1.5221	20.0	3800	1.9030	43.2794	25.4126	39.2902	39.4092
1.5221	21.0	3990	1.8996	43.1731	25.3819	39.1873	39.3172
1.5221	22.0	4180	1.9006	43.4949	25.4485	39.3092	39.4516
1.4714	23.0	4370	1.8977	43.5657	25.5974	39.4489	39.5257
1.4714	24.0	4560	1.9035	43.6444	25.6794	39.5809	39.683
1.4421	25.0	4750	1.9000	43.4825	25.5898	39.4319	39.4973
1.4421	26.0	4940	1.9030	43.4623	25.5726	39.461	39.6009
1.4421	27.0	5130	1.8993	43.3357	25.5518	39.3897	39.4672
1.4139	28.0	5320	1.9009	43.5834	25.7211	39.584	39.6725
1.4139	29.0	5510	1.9002	43.7115	25.6997	39.6603	39.7621
1.4016	30.0	5700	1.9013	43.7595	25.7022	39.6153	39.7151

Framework versions

• Transformers 4.25.1
• Pytorch 1.13.0
• Datasets 2.8.0
• Tokenizers 0.12.1